Customizing CLM history fields (CLM4)

[fsche@mech_kth_se]

Hi,I tried to add several history fields for output in my CESM1.0.5 (F_1850, CAM4, CLM4) simulation, but it doesn't work when changing clm.buildnml.csh My namelist file looks like this:&clm_inparm
 co2_type               = 'diagnostic'
 create_crop_landunit           = .false.
 dtime          = 1800
 fatmgrid        ....  urban_traffic          = .false.
 hist_nhfrq          = 0,-24,-3,-1
 hist_mfilt          = 1,30,240,720
 hist_fincl1         = 'AGDD','ALBD'
 hist_fincl2         = 'FSNO','H2OSNO','PLAI','SNOWDP','TLAI','TSA:X','TSA:M','TG:X','TG:M'
 hist_fincl3         = 'FIRA','PET','Q2M','Qle','QMELT','QOVER','QSOIL','QVEGE','QVEGT',
                                  'RAIN','RH2M','Rnet','TG','TLAKE','TSA','TSOI','TSOI_10CM','U10','WIND'
 hist_fincl4         = 'SNOW'
/
&ndepdyn_nml

I selected the variables above according to the CLM history fields table:http://www.cesm.ucar.edu/models/cesm1.0/clm/models/lnd/clm/doc/UsersGuide/x1997.html Does anybody see an error there? The lnd.logfile complains about it at least, but I don't see any error.lnd.log.141106-203349 yields(lnd_init_mct) :CLM land model initialization
 cesm1_0_5

 Attempting to initialize the land model .....

 Attempting to initialize run control settings .....
 Read in clm_inparm namelist from: lnd_in

 

[slevis]

Looks right syntactically as far as I can tell, though PLAI and Qle do not ring a bell. You may wish to look in CLM's histFldsMod.F90, in case some of these variables are inactive or not present. If this does not resolve the issue, then I suggest modifying the namelist a little bit at a time until you bump up against the offending variable that causes the crash.Sam Levis

 

[fsche@mech_kth_se]

Thank you very much for your reply. Above was a typing error: a missing t for hist_nhtfrq, but that was not the main issue.You're right, the variable list linked above is not the master field for the CLM output. I checked now the default master field that is written out during runtime. There are only 239 variables in the lnd-logfile. So some variables I needed are just not defined.However, when looking in the file you suggested, histFldsMod.F90 includes all variables I wanted to write out but not all of them are in the master field during run time. So I don't understand how I can write out variables like "PET" for "Potential Evapotranspiration" or "ALBD" for "surface albedo (direct)" etc.The odd thing is that e.g. Qle and Rnet are default='inactive' in the histFldsMod.F90 but still in the master field while PET has no flag for inactive but is not in the master field. PET and AGDD have been within a loop "if(defined CASA)" or "if (defined CNDV)", which were both commented out. Not sure what is going on there.