3. Generate the Workflow Experiment

3. Generate the Workflow Experiment
schramm Thu, 08/12/2021 - 08:35

Estimated time: 15 minutes

Generating the workflow requires three steps:

  • Set experiment parameters in config.sh
  • Set Python and other environment parameters
  • Run the generate_FV3LAM_wflow.sh script

The workflow requires a file called config.sh to specify the values of your experiment parameters. Make a copy of the example template config.community.sh file provided in the $SR_WX_APP_TOP_DIR/regional_workflow/ush directory (from the build directory):

cd ../regional_workflow/ush
cp config.community.sh config.sh

Edit the config.sh file on Cheyenne for running the FV3_GFS_v15p2 physics suite for 48 hours starting at 00 UTC on 15 June 2019 on a 25 km predefined CONUS domain.  Since this is the default case, only the following variables need to be modified in config.sh:

MACHINE="cheyenne"
ACCOUNT="NJNT0008"
USE_USER_STAGED_EXTRN_FILES="TRUE"
EXTRN_MDL_SOURCE_BASEDIR_ICS="/glade/p/ral/jntp/UFS_SRW_app/staged_extrn_mdl_files/FV3GFS"
EXTRN_MDL_SOURCE_BASEDIR_LBCS="/glade/p/ral/jntp/UFS_SRW_app/staged_extrn_mdl_files/FV3GFS"

where ACCOUNT is an account you can charge. We will be using pre-staged initialization data, so uncomment USE_USER_STAGED_EXTRN_FILES="TRUE".

EXTRN_MDL_SOURCE_BASEDIR_ICS and EXTRN_MDL_SOURCE_BASEDIR_LBCS contain the locations of the initial and lateral boundary conditions, and EXTRN_MDL_FILES_ICS and EXTRN_MDL_FILES_LBCS contain the filenames of the initial and lateral boundary conditions.

Load the appropriate Python environment for Cheyenne:

source ../../env/wflow_cheyenne.env

If you see something like:

Warning: library /glade/p/ral/jntp/UFS_SRW_app/ncar_pylib/regional_workflow is missing from NPL clone registry.
Do you wish to add it to the registry (y/N)?

This will occur if this is the first time you have used the NCAR Python Library (NPL), type y, to get the following output:

Searching library for configuration info ...
Now using NPL virtual environment at path:
/glade/p/ral/jntp/UFS_SRW_app/ncar_pylib/regional_workflow
Use deactivate to remove NPL from environment

The workflow can be generated with the command:

./generate_FV3LAM_wflow.sh

The generated workflow will be located in ${SR_WX_APP_TOP_DIR}/../expt_dirs/${EXPT_SUBDIR} where ${SR_WX_APP_TOP_DIR} is the directory where the ufs-srweather-app was cloned in Step 1, and ${EXPT_SUBDIR} is the experiment name set above in the config.sh file: test_CONUS_25km_GFSv15p2