Error in run, pio_support::pio_die:: myrank= -1 : ERROR: nf_mod.F90: 1288 : NetCDF: Invalid dimension ID or name

[zpc]

the follow is my code
./create_newcase --case ~/cases/preheat/pft1_1 --res CLM_USRDAT --compset 2000_DATM%GSWP3v1_CLM50%BGC_SICE_SOCN_SROF_SGLC_SWAV --run-unsupported --compiler intel --mach myintel


# Change accelerated spinup mode
./xmlchange CLM_ACCELERATED_SPINUP="on"

./xmlchange CLM_USRDAT_NAME='tibet'
./xmlchange ATM_DOMAIN_PATH='/home/user/own.roi/myout_1grid_1/'
./xmlchange LND_DOMAIN_PATH='/home/user/own.roi/myout_1grid_1/'
./xmlchange ATM_DOMAIN_FILE='domain.lnd.fv0.23x0.31_tx0.1v2_070929_tibet_pt1_1.nc'
./xmlchange LND_DOMAIN_FILE='domain.lnd.fv0.23x0.31_tx0.1v2_070929_tibet_pt1_1.nc'

./xmlchange NTASKS=1

./xmlchange STOP_N=1
./xmlchange STOP_OPTION=nyears
./xmlchange RESUBMIT=99

./xmlchange DATM_CLMNCEP_YR_START=2000
./xmlchange DATM_CLMNCEP_YR_END=2000
./xmlchange DATM_CLMNCEP_YR_ALIGN=1
./xmlchange RUN_REFDATE=0001-01-01
./xmlchange RUN_STARTDATE=0001-01-01

./case.setup --reset

gedit ./user_nl_clm

echo "fsurdat = '/home/user/own.roi/myout_1grid_1/surfdata_0.23x0.31_simyr2000_c100406_tibet_pt1_1.nc'
hist_empty_htapes = .true.
hist_fincl1='GPP:A','NPP:A','NEP:A','NEE:A','NBP:A','TOTECOSYSC:A','TOTSOMC:A','TOTVEGC:A','TLAI:A','TWS:A','H2OSNO:A'
hist_fincl2='GPP:X','NPP:X','NEP:X','NEE:X','NBP:X','TOTECOSYSC:X','TOTSOMC:X','TOTVEGC:X','TLAI:X','TWS:X','H2OSNO:X'
hist_nhtfrq=0,0 !monthly average
hist_mfilt =3600,3600
use_init_interp = .true." >> ./user_nl_clm

./case.build

./case.submit

 

[zpc]

the follow is my code files

 

[oleson]

The traceback in the cesm log indicates:

cesm.exe 0000000000DBF790 surfrdmod_mp_surf 514 surfrdMod.F90

Line 514 in the release-cesm2.1.0 version of surfrdMod.F90 is:

call check_dim(ncid, 'nglcecp1', maxpatch_glcmec+1 )

So that implies you don't have that dimension (nglcecp1) on your surface dataset. That dimension is associated with variable GLC_MEC on the surface dataset. Do you have that variable on your dataset file?
On the other hand, I don't see where GLC_MEC is read in or used, at least in release-cesm2.1.0, and I don't see nglcecp1 used or read in anywhere either except for that one line. So I'm wondering if you can just comment out that one line and rebuild the code.

 

[zpc]

The traceback in the cesm log indicates:

cesm.exe 0000000000DBF790 surfrdmod_mp_surf 514 surfrdMod.F90

Line 514 in the release-cesm2.1.0 version of surfrdMod.F90 is:

call check_dim(ncid, 'nglcecp1', maxpatch_glcmec+1 )

So that implies you don't have that dimension (nglcecp1) on your surface dataset. That dimension is associated with variable GLC_MEC on the surface dataset. Do you have that variable on your dataset file?
On the other hand, I don't see where GLC_MEC is read in or used, at least in release-cesm2.1.0, and I don't see nglcecp1 used or read in anywhere either except for that one line. So I'm wondering if you can just comment out that one line and rebuild the code.

thank you for your reply!
yes, l dont need this variable. In fact, l just need to clip a surfdata to run my region. the fsurdata version is surfdata_0.23x0.31_simyr2000_c100406.nc, but the version is too old, lacking some variable. do you have any new version 0.23x0.31 surfdata?
On the other hand, if l comment out that one line, l will get some others errors.
thank you so much!

 

[oleson]

I think you should be able to create one using mksurfdata_map. See the general instructions here:

1.3.5. Creating Surface Datasets — ctsm release-clm5.0 documentation

More simply, this worked for me on our HPC machine using release-cesm2.1.3 in components/clm/tools/mksurfdata_map:

qcmd -- ./mksurfdata_map.pl -res 0.23x0.31

The qcmd command submits the job to our HPC compute nodes.