Trouble Running CLM-FATES and spinup

[LYI]

Hello everyone,
I am running CESM2.2 with CLM5 + FATES, and my workflow has two steps:
(1) Spin-up case
Goal: cold-start spin-up with compset 2000_DATM%GSWP3v1_CLM50%FATES_SICE_SOCN_SROF_SGLC_SWAV, resolution 1x1_brazil.
Key xmlchange settings:
RUN_STARTDATE=2010-01-01
STOP_N=30
STOP_OPTION=ndays
REST_N=5,REST_OPTION=ndays
DATM_CLMNCEP_YR_START=2010
DATM_CLMNCEP_YR_END=2010
DATM_CLMNCEP_YR_ALIGN=2010
DOUT_S=FALSE
CLM_FORCE_COLDSTART=on
NTASKS=1
./xmlchange LND_DOMAIN_PATH=/inputdata/share/domains/domain.clm
./xmlchange LND_DOMAIN_FILE=domain.lnd.1x1pt-brazil_navy.090715.nc
./xmlchange ATM_DOMAIN_PATH=/inputdata/share/domains/domain.clm
./xmlchange ATM_DOMAIN_FILE=domain.lnd.1x1pt-brazil_navy.090715.nc
user_nl_clm:
use_fates = .true.
fsurdat = '/inputdata/lnd/clm2/surfdata_map/release-clm5.0.18/surfdata_1x1_brazil_hist_16pfts_Irrig_CMIP6_simyr2000_c190214.nc'
paramfile = '/inputdata/lnd/clm2/paramdata/clm5_params.c200624.nc'
Question 1:
Are these settings correct and internally consistent with this compset/res for a cold-start spin-up?

(2)CLM-FATES run using spin-up result（finidat_spinup0811_20100131.nc）as finidat
Key xmlchange settings:
./xmlchange LND_DOMAIN_PATH=/inputdata/share/domains/domain.clm
./xmlchange LND_DOMAIN_FILE=domain.lnd.1x1pt-brazil_navy.090715.nc
./xmlchange ATM_DOMAIN_PATH=/inputdata/share/domains/domain.clm
./xmlchange ATM_DOMAIN_FILE=domain.lnd.1x1pt-brazil_navy.090715.nc
./xmlchange RUN_STARTDATE=2010-01-01
./xmlchange DATM_CLMNCEP_YR_START=2010
./xmlchange DATM_CLMNCEP_YR_END=2010
./xmlchange DATM_CLMNCEP_YR_ALIGN=2010
./xmlchange STOP_OPTION=ndays
./xmlchange STOP_N=30
user_nl_clm:
use_fates = .true.
fates_parteh_mode = 1
fsurdat = '/inputdata/lnd/clm2/surfdata_map/release-clm5.0.18/surfdata_1x1_brazil_hist_16pfts_Irrig_CMIP6_simyr2000_c190214.nc'
paramfile = '/inputdata/lnd/clm2/paramdata/clm5_params.c200624.nc'
finidat = '/inputdata/lnd/clm2/initdata_map/finidat_spinup0811_20100131.nc'
hist_nhtfrq = -24
hist_mfilt = 1
hist_fincl2 = 'FATES_MORTALITY', 'FATES_MORT_RATE_HYDRO', 'FATES_MORT_RATE_CSTARV', 'FATES_COHORT_MORTALITY_RATE', 'FATES_LEAF_WATER_POTENTIAL', 'FATES_BALIVE', 'FATES_LEAF_AREA_INDEX'

When running this case, I get error1(.txt)
Question 2:
Is this segmentation fault more likely due to an issue in the finidat contents itself, or due to incompatibility from changing compset/grid/build configurations. And how can I identify the source of the problem?

 

[oleson]

If you removed the following from user_nl_clm in your second case does it run?

hist_nhtfrq = -24
hist_mfilt = 1
hist_fincl2 = 'FATES_MORTALITY', 'FATES_MORT_RATE_HYDRO', 'FATES_MORT_RATE_CSTARV', 'FATES_COHORT_MORTALITY_RATE', 'FATES_LEAF_WATER_POTENTIAL', 'FATES_BALIVE', 'FATES_LEAF_AREA_INDEX'

 

[LYI]

If you removed the following from user_nl_clm in your second case does it run?

hist_nhtfrq = -24
hist_mfilt = 1
hist_fincl2 = 'FATES_MORTALITY', 'FATES_MORT_RATE_HYDRO', 'FATES_MORT_RATE_CSTARV', 'FATES_COHORT_MORTALITY_RATE', 'FATES_LEAF_WATER_POTENTIAL', 'FATES_BALIVE', 'FATES_LEAF_AREA_INDEX'

Thank you! I tried removing the following settings from user_nl_clm as you suggested. The case ran successfully, but I only got the following output files:
case_test_fates.clm2.r.2010-01-11-00000.nc
case_test_fates.cpl.r.2010-01-11-00000.nc
case_test_fates.clm2.rh0.2010-01-11-00000.nc

 

[oleson]

You only ran 30 days, you would need to run 31 days (January) to get a monthly history file (clm2.h0).
I see you wanted daily average values of the variables listed in hist_fincl2. But I'm not sure some of those variables are available. For example, I don't see a FATES_MORTALITY variable in the code, I do see a MORTALITY variable. Another example is that I don't see a FATES_BALIVE variable (or a BALIVE) in the code. I'm assuming the error you got could possibly be associated with requesting non-existent variables, since your run is successful if you don't specify those variables. Where did your list come from?

 

[LYI]

You only ran 30 days, you would need to run 31 days (January) to get a monthly history file (clm2.h0).
I see you wanted daily average values of the variables listed in hist_fincl2. But I'm not sure some of those variables are available. For example, I don't see a FATES_MORTALITY variable in the code, I do see a MORTALITY variable. Another example is that I don't see a FATES_BALIVE variable (or a BALIVE) in the code. I'm assuming the error you got could possibly be associated with requesting non-existent variables, since your run is successful if you don't specify those variables. Where did your list come from?

Thank you for your guidance. After adjusting the simulation period, I was able to get the h0 files. I realized that some of my output variable requests were invalid. I found some variable names in the CTSM user guide such as 'FATES_MORTALITY_CSTARV_CFLUX_PF' and 'FATES_MORTALITY_CFLUX_PF', but it seems these are still not valid in my version. Could you please point me to any documentation or references for the available history variables for CLM5.0 + FATES in CESM2.2?

 

[LYI]

I also tried to look for 'FATES_MORTALITY' related variables in the model by checking main/FatesHistoryInterfaceMod.F90 as indicated in the FATES Tutorial. However, it seems that these variables are still not valid in my CESM2.2 + CLM5.0 setup.

 

[oleson]

I don't see a list of FATES variables available in CLM5-FATES in the User's Guide unfortunately. Just for the CTSM development version here: 1.2.6. CTSM History Fields (fates) — ctsm CTSM master documentation. You could post in the FATES discussion board to see if such a list is available: NGEET fates · Discussions

 

[LYI]

I don't see a list of FATES variables available in CLM5-FATES in the User's Guide unfortunately. Just for the CTSM development version here: 1.2.6. CTSM History Fields (fates) — ctsm CTSM master documentation. You could post in the FATES discussion board to see if such a list is available: NGEET fates · Discussions

Thank you very much for your reply and for sharing the link! I’ll check the FATES discussion board to see if there’s a more complete list of available variables.