No output for case "fv3_model_point_noah" (v6.0.0)

[wzzheng90]

Hi, I am trying to test v6.0.0 with the provided cases on HERA. I got a problem for running case "fv3_model_point_noah" (with all default config).
After running it: ./run_scm.py -c fv3_model_point_noah -s SCM_GFS_v16 , I did not find the output file "output.nc" under ~/run/output_fv3_model_point_noah_SCM_GFS_v16. There are only two files "fv3_model_point_noah_SCM_GFS_v16.nml" and "logfile" under this directory. Also, I did not find any useful information from this "logfile". Anything I did not set up correctly? Thanks! -- Weizhong

[grantfirl]

Hi Weizhong,

It looks like the input data that the physics expects has changed to the point that what the UFS_IC_generator.py script no longer is providing the right data. We will need to use the latest FV3GFS_io.F90 file in the fv3atm repository together with up-to-date UFS initial condition files to update the UFS_IC_generator.py script to regenerate a new fv3_model_point_noah case data file. Let me discuss with the DTC team about how/if we can accomplish this with our deliverables/funding.

-Grant

[wzzheng90]

Hi, Grant,

Thanks for your quick response. Looking forward to the updated code/scripts.
-Weizhong

[christianbsch]

Hello

I am having the same problem, no output file for the test case fv3_model_point_noah... there is any update about the FV3GFS_io.F90 and the UFS_IC_generator.py script? This latter only generates Initial conditions, but not forcing data...

Thanks in advance for your help

Regards

Christian

[dustinswales]

Hi Christan,

Sorry you are having troubles.
The script in question has been updated. Please switch to the main branch and refer to the documentation

[christianbsch]

Hello Dustin, thanks for the reply. I have used the scripts from the main branch, however:

1. The fv3_model_point_noah experiment still doesn't generate any output
2. I tried to create again the fv3_model_point_noah netcdf file with UFS_IC_generator.py (with the main branch version), but it shows the following error: " UFS_IC_generator.py: error: the following arguments are required: -f/--forcing_dir ",
   Thank you

[ulmononian]

i was wondering if there was an update on this?

on hera, i did the following

git clone --recursive -b v6.0.0 https://github.com/NCAR/ccpp-scm
cd ccpp-scm
source scm/etc/Hera_setup_intel.sh
./contrib/get_all_static_data.sh
./contrib/get_thompson_tables.sh
cd scm
mkdir bin; cd bin
cmake ../src 2>&1 | tee log.cmake.001
make -j8 2>&1 | tee log.make.001
./run_scm.py -m 2>&1 | tee log.run_scm.001

perhaps i misunderstood, but i expected all suites & cases to run with ./run_scm.py -m, but there is no output folder created for fv3_model_point_noah generated in ccpp-scm/scm/run. there are output folders here for several others cases (e.g. arm, bomex, LASSO, astex) but not the ufs case.

when i then try just ./run_scm.py -c fv3_model_point_noah -s SCM_GFS_v16, i get the following error

(pyccpp) [Cameron.Book@hfe01 bin]$ ./run_scm.py -c fv3_model_point_noah -s SCM_GFS_v16
INFO: Setting up experiment fv3_model_point_noah with suite SCM_GFS_v16 using namelist input_GFS_v16.nml and tracers tracers_GFS_v16.txt
INFO: Parsing case configuration namelist /scratch1/NCEPDEV/stmp4/Cameron.Book/835/scm-v6/scm/etc/../../scm/etc/case_config/fv3_model_point_noah.nml
INFO: Creating experiment configuration namelist input_experiment.nml in the run directory from /scratch1/NCEPDEV/stmp4/Cameron.Book/835/scm-v6/scm/etc/../../scm/etc/case_config/fv3_model_point_noah.nml using SCM_GFS_v16 and input_GFS_v16.nml
INFO: Launching experiment fv3_model_point_noah with suite SCM_GFS_v16 using namelist input_GFS_v16.nml and tracers tracers_GFS_v16.txt
INFO: Passing control to "(cd /scratch1/NCEPDEV/stmp4/Cameron.Book/835/scm-v6/scm/etc/../../scm/run && time /scratch1/NCEPDEV/stmp4/Cameron.Book/835/scm-v6/scm/etc/../../scm/run/scm)"
CRITICAL: Execution of command "(cd /scratch1/NCEPDEV/stmp4/Cameron.Book/835/scm-v6/scm/etc/../../scm/run && time /scratch1/NCEPDEV/stmp4/Cameron.Book/835/scm-v6/scm/etc/../../scm/run/scm)" failed, exit code 127
    stdout: ""
    stderr: "/scratch1/NCEPDEV/stmp4/Cameron.Book/835/scm-v6/scm/etc/../../scm/run/scm: error while loading shared libraries: libzstd.so.1: cannot open shared object file: No such file or directory

real	0m0.070s
user	0m0.001s
sys	0m0.015s"
Error in atexit._run_exitfuncs:
Traceback (most recent call last):
  File "./run_scm.py", line 733, in launch_executable
    (status, stdout, stderr) = execute(cmd, ignore_error = ignore_error)
  File "./run_scm.py", line 173, in execute
    raise Exception('Execution of command "{0}" failed, exit code {1}\n'.format(cmd, status))
Exception: Execution of command "(cd /scratch1/NCEPDEV/stmp4/Cameron.Book/835/scm-v6/scm/etc/../../scm/run && time /scratch1/NCEPDEV/stmp4/Cameron.Book/835/scm-v6/scm/etc/../../scm/run/scm)" failed, exit code 127

do i need to re-do the compile steps to invoke ./run_scm.py -c fv3_model_point_noah -s SCM_GFS_v16 or should this work from the initial build steps i shared at the top of this message?

any suggestions or insight would be most appreciated!

[wzzheng90]

@ulmononian I got the similar error for this case with the latest SCM version. The output folder was created for fv3_model_point_noah (generated in ccpp-scm/scm/run), but NO 'output.nc" was generated. Looking forward to any suggestions/solutions too. Thanks.
-- Weizhong

[ulmononian]

i just repeated the instructions i shared in my first post, but using the ccpp-scm master branch. this time, ./run_scm.py -c fv3_model_point_noah -s SCM_GFS_v16 runs successfully. however, similar to what @wzzheng90 mentioned, an output_fv3_model_point_noah_SCM_GFS_v16 was created in run but with no output.nc file (just a .nml and logfile).

any suggestions?

[hertneky]

We are currently looking into a solution and will respond as soon as we have something ready for you. Sorry for the inconvenience. Bottom line, the canned fv3_model_point_noah will not work and we are working on getting replay mode ready for you to generate the UFS forcing files instead. More to come soon.

[ulmononian]

@hertneky thank you very much for looking into this!

[wzzheng90]

@hertneky Any updates about this issue? Thanks!