pio_support::pio_die:: myrank= -1 : ERROR: ionf_mod.F90: 235 : Invalid argument

[Li Ming]

Hi,
The version I am using is release-CLM5.0.35. I want to researching the NPP in the historical period with my own 0.1° atmospheric forcing data combined with a 0.1° domain that I created for my study area. The command I used to create the case was ./create_newcase --case test9 --res CLM_USRDAT --compset HIST_DATM%GSWP3v1_CLM50%BGC_SICE_SOCN_SROF_SGLC_SWAV --compiler gnu --mach mygnu. In user_nl_clm I added the finidat file path, but the model reports the following error:

---

NetCDF: Invalid dimension ID or name
NetCDF: Invalid dimension ID or name
NetCDF: Invalid dimension ID or name
NetCDF: Invalid dimension ID or name
NetCDF: Invalid dimension ID or name
NetCDF: Variable not found
NetCDF: Variable not found
Invalid argument
pio_support::pio_die:: myrank= -1 : ERROR: ionf_mod.F90: 235 : Invalid argument
Abort(1) on node 1 (rank 1 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, 1) - process 1

---

When I delete my own finidat file path in user_nl_clm and use the default finidat like clmi.I2000Clm50BgcCrop.2011-01-01.1.9x2.5_gx1v7_gl4_simyr2000_c190312.nc , the model is running perfectly fine. I thought this would indicate that my atmospheric forcing data is fine, so I checked my finidat and found that I used a surface file similar to 2000 years in doing the SPIN UP process, instead of a surface similar to 1850 which is needed for HIST case, is the error reported because of this?

Any help will be much appreciated. Thanks!

 

[Li Ming]

Sorry, I have another question, since the time period I am researching is 1979-2018, I would like to run the HIST case with surfdata_CMFD-China_hist_16pfts_Irrig_CMIP6_simyr2000_c230903.nc, but this requires that the similar year of the surfaces match the landuse_timeseries, can I make a landuse_timeseries file close to 2000 to run the HIST case? If so, how do I need to do this, as the landuse_timeseries files I have made myself are all close to 1850
Thanks for your reply.

 

[oleson]

I'm not sure you are specifying the finidat file correctly. In lnd_in, it seems to be:

finidat = '/gpfs/gpfs_fs/home/xmx/clm5new/cime/scripts/ice-10'

Is this simply the path to the file and you've not included the actual file?

 

[Li Ming]

Thankyou for your reply. I ignored this part and the model ran fine after entering the full finidat.
I have a new problem, I have successfully made the domain file and lower resolution dataset surface file, but I want to make the high-resolution pft dataset suface file, I enter the following command

---

/mksurfdata.pl -r usrspec -usr_gname $GRIDNAME -usr_gdate $CDATE -l $CSMDATA -years 2000 -hirespft

---

there is an error and no information after executing DefaultNamelist.pl according to the error message. How can I solve this problem?

 

[oleson]

Try -years 2005. I think the hirespft is keyed off of that year.

 

[Houhhu]

Sorry, I have another question, since the time period I am researching is 1979-2018, I would like to run the HIST case with surfdata_CMFD-China_hist_16pfts_Irrig_CMIP6_simyr2000_c230903.nc, but this requires that the similar year of the surfaces match the landuse_timeseries, can I make a landuse_timeseries file close to 2000 to run the HIST case? If so, how do I need to do this, as the landuse_timeseries files I have made myself are all close to 1850
Thanks for your reply.

@Li Ming Hi, I met the same problem about

the similar year of the surfaces match the landuse_timeseries.

I would like to run the HIST compset with my own surfdata and landuse_timeseires files from 1983 to 2015, but the time period of landuse_timeseries data is 1850 to 2015. The error in dynpftFileMod.F90 occurs when I submit the job.
"dynpft_check_consistency mismatch between PCT_NAT_PFT at initial time and that obtained from surface dataset"

 

[slevis]

In user_nl_clm you can try "check_dynpft_consistency = .false."

A recent version of the model that I'm looking at would have mentioned this in the error message that you got.

 

[Houhhu]

In user_nl_clm you can try "check_dynpft_consistency = .false."

A recent version of the model that I'm looking at would have mentioned this in the error message that you got.

Thanks for you respone, slevis. I have seen this "check_dynpft_consistency = .false." before the error message.Just want to confirm that the time of landuse_timeseries data read would be consistent with the model RUN_STARTDATE.
eg. model starts at 1983-01-01, landuse_timeserise data read at time 1983.

 

[slevis]

If you want the simulation to use 1983-2015 landuse data, then set
RUN_STARTDATE = 1983-01-01
DATM_YR_ALIGN = 1983
DATM_YR_START = 1983
DATM_YR_END = 2015

It's probably ok that your landuse_timeseries spans 1850-2015, as long as it contains the modified file paths from 1983-2015 to point to your custom-made files.

 

[Houhhu]

Thanks for your response. I know what to do now.

 

[Houhhu]

Dear Slevis,
I run the model with my own forcing from 1983 to 2016 with HIST compset. But I notice the CO2 file the model used is "fco2_datm_global_simyr_1750_2014_CMIP6_c180929.nc". It seems the file doesn't have the CO2 data in 2015 and 2016.
I set the taxmode "extend " for CO2 file. The model runs successfully, but I cannot find CO2 file read by the model after model year 2014 in atm.log. Will the model use CO2 data in the last two years 2015 and 2016?
looking forward to you reply,thanks!

 

[slevis]

I think "extend" means that the model will keep using the last value that it read, in your case the 2014 value.