error with the customized land use data

[xgao304]

Dear Sir/Madam,

I am using cesm2.1.3 to run the following compset: 2000_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_SROF_SGLC_SWAV
I am testing two runs with my customized land use time series data and forcing data from 2021 to 2050:

1. CLM5 run from 2021 to 2050 with the static land use data of 2021
2. CLM5 run from 2021 to 2050 with the dynamic land use data of 2021 - 2050

Currently, case 1 has been running successfully, but the case 2 cannot run. The error messages are not evident (please see the attached log files).
I am creating the case 2 with the following command:

create_newcase --case BgcCrop2000_USA_demeterEPPA_dyn2021
--compset 2000_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_SROF_SGLC_SWAV \
--res hcru_hcru --user-mods-dir $MYDATA_DIR --machine svante --compiler intel \
--run-unsupported

Any idea about what is going wrong?

Thanks,

Xiang

 

[Yuan Sun]

Dear Sir/Madam,

I am using cesm2.1.3 to run the following compset: 2000_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_SROF_SGLC_SWAV
I am testing two runs with my customized land use time series data and forcing data from 2021 to 2050:

1. CLM5 run from 2021 to 2050 with the static land use data of 2021
2. CLM5 run from 2021 to 2050 with the dynamic land use data of 2021 - 2050

Currently, case 1 has been running successfully, but the case 2 cannot run. The error messages are not evident (please see the attached log files).
I am creating the case 2 with the following command:

create_newcase --case BgcCrop2000_USA_demeterEPPA_dyn2021
--compset 2000_DATM%GSWP3v1_CLM50%BGC-CROP_SICE_SOCN_SROF_SGLC_SWAV \
--res hcru_hcru --user-mods-dir $MYDATA_DIR --machine svante --compiler intel \
--run-unsupported

Any idea about what is going wrong?

Thanks,

Xiang

Hi Xiang,

The cesm.log noted that NetCDF: Invalid dimension ID or name, suggesting something is wrong with your land inputs.

The lnd.log noted that:
Reading in urban input data from fsurdat file ...
(GETFIL): attempting to find local file
landuse.timeseries_360x720cru_BAU_78pfts_USA_demeter_EPPA_dyn_2021_2050.nc
(GETFIL): using
/net/fs05/d1/xgao/cesm2.1.3/cases/sim_setup/demeter_EPPA/USA_surfdomain_dyn/lan
duse.timeseries_360x720cru_BAU_78pfts_USA_demeter_EPPA_dyn_2021_2050.nc
UrbanInput
/net/fs05/d1/xgao/cesm2.1.3/cases/sim_setup/demeter_EPPA/USA_surfdomain_dyn/lan
duse.timeseries_360x720cru_BAU_78pfts_USA_demeter_EPPA_dyn_2021_2050.nc

This indicated that you set landuse data to fsurdat. You need to point to a correct fsurdat file.

Best,
Yuan

 

[xgao304]

@Yuan,

Thanks for your reply. I am not sure if NetCDF: Invalid dimension ID or name is directly linked to an error with land inputs. For the successful run, this kind of message is also found in the cesm.log file.

The file is there (see the attached image):

Thanks,

Xiang

 

[Yuan Sun]

@Yuan,

Thanks for your reply. I am not sure if NetCDF: Invalid dimension ID or name is directly linked to an error with land inputs. For the successful run, this kind of message is also found in the cesm.log file.

The file is there (see the attached image):

Thanks,

Xiang

In my understanding, transient simulations require two inputs.

For example, the transient job script is:
export SURF='XXX'
export TRANSURF='/net/fs05/d1/xgao/cesm2.1.3/cases/sim_setup/demeter_EPPA/USA_surfdomain_dyn/landuse.timeseries_360x720cru_BAU_78pfts_USA_demeter_EPPA_dyn_2021_2050.nc'
echo "fsurdat='${SURF}'" >> user_nl_clm
echo "flanduse_timeseries='${TRANSURF}'" >> user_nl_clm

The static job script is:
export SURF='XXX'
export TRANSURF=' '
echo "fsurdat='${SURF}'" >> user_nl_clm
echo "flanduse_timeseries='${TRANSURF}'" >> user_nl_clm

Best,
Yuan

 

[xgao304]

Thanks for the information. For the static job, I have not specified flanduse_timeseries in "user_nl_clm", and the run seems going fine.
But I will try to add flanduse_timeseries for the transient run and give it a try. I have one question: If I specify both fsurdat and flanduse_timeseries,
would there be any conflict information? For example, fsurdata generally contains percentages of land unit, pft, cft (assuming those are all fixed) within each grid, then flanduse_timeseries may also contain similar information but it will vary annually (dynamic).

Thanks,

Xiang

 

[Yuan Sun]

Thanks for the information. For the static job, I have not specified flanduse_timeseries in "user_nl_clm", and the run seems going fine.
But I will try to add flanduse_timeseries for the transient run and give it a try. I have one question: If I specify both fsurdat and flanduse_timeseries,
would there be any conflict information? For example, fsurdata generally contains percentages of land unit, pft, cft (assuming those are all fixed) within each grid, then flanduse_timeseries may also contain similar information but it will vary annually (dynamic).

Thanks,

Xiang

In my understanding, at the initialization stage, CTSM gets constant parameters from the fsurfdata first.

Then, transient simulations get the landunit fraction (PCT_*) for a certain year (according to the model year) from flanduse_timeseries during the computation.