1.1.2. Technical Overview

This chapter provides information on SRW App prerequistes, component code repositories, and SRW App directory structure.

1.1.2.1. Prerequisites for Using the SRW Application

1.1.2.1.1. Background Knowledge Prerequisites

The instructions in this documentation assume that users have certain background knowledge:

Familiarity with LINUX/UNIX systems
Command line basics
System configuration knowledge (e.g., compilers, environment variables, paths, etc.)
Numerical Weather Prediction (e.g., concepts of parameterizations: physical, microphysical, convective)
Meteorology (in particular, meteorology at the scales being predicted: 25km, 13km, and 3km resolutions)

Additional background knowledge in the following areas could be helpful:

High-Performance Computing (HPC) Systems (for those running the SRW App on an HPC system)
Programming (particularly Python and bash scripting) for those interested in contributing to the SRW App code
Creating an SSH Tunnel to access HPC systems from the command line
Containerization
Workflow Managers/Rocoto

1.1.2.1.2. Software/Operating System Requirements

The UFS SRW Application has been designed so that any sufficiently up-to-date machine with a UNIX-based operating system should be capable of running the application. SRW App Level 1 & 2 systems already have these prerequisites installed. However, users working on other systems must ensure that the following requirements are installed on their system:

Minimum Platform Requirements:

POSIX-compliant UNIX-style operating system
>97 GB disk space
- 53 GB input data for a standard collection of global data, or “fix” file data (topography, climatology, observational data) for a short 12-hour test forecast on the CONUS 25km domain. See data download instructions in Section 1.3.2.3.
- ~23 GB for full spack-stack installation (or ~8 GB HPC-Stack)
- 3 GB for ufs-srweather-app installation
- 1 GB for boundary conditions for a short 12-hour test forecast on the CONUS 25km domain. See data download instructions in Section 1.3.2.3.
- 17 GB for a 12-hour test forecast on the CONUS 25km domain, with model output saved hourly.
Fortran compiler released since 2018
- gfortran v9+ or ifort v18+ are the only ones tested, but others may work.
C and C++ compilers compatible with the Fortran compiler
- gcc v9+, ifort v18+, and clang v9+ (macOS, native Apple clang, LLVM clang, GNU) have been tested
Python v3.7+ (preferably 3.9+), including prerequisite packages jinja2, pyyaml, and f90nml
- Python packages scipy, matplotlib, pygrib, cartopy, and pillow are required for users who would like to use the provided graphics scripts.
Perl 5
git v2.12+
Lmod
wget
- Only required for retrieving data using retrieve_data.py. If data is prestaged, wget is not required. If data is retrieved using other means, curl may be used as an alternative.

The following software is also required to run the SRW Application, but the spack-stack (which contains the software libraries necessary for building and running the SRW App) can be configured to build these requirements:

CMake v3.20+
MPI (MPICH, OpenMPI, or other implementation)
- Only MPICH or OpenMPI can be built with spack-stack. Other implementations must be installed separately by the user (if desired).

For MacOS systems, some additional software packages are needed. When possible, it is recommended that users install and/or upgrade this software (along with software listed above) using the Homebrew package manager for MacOS. See HPC-Stack Documentation: Chapter 3 for further guidance on installing these prerequisites on MacOS.

bash v4.x
GNU compiler suite v11 or higher with gfortran
cmake
make
coreutils
gsed

Optional but recommended prerequisites for all systems:

Conda for installing/managing Python packages
Bash v4+
Rocoto Workflow Management System (1.3.1)
Python packages scipy, matplotlib, pygrib, cartopy, and pillow for graphics

1.1.2.2. Code Repositories and Directory Structure

1.1.2.2.1. Hierarchical Repository Structure

The umbrella repository for the SRW Application is named ufs-srweather-app and is available on GitHub at https://github.com/ufs-community/ufs-srweather-app. The SRW Application uses the manage_externals tool and a configuration file called Externals.cfg to pull in the appropriate versions of the external repositories associated with the SRW App (see Table 1.2).

Table 1.2 List of top-level repositories that comprise the UFS SRW Application
Repository Description	Authoritative repository URL
Umbrella repository for the UFS Short-Range Weather (SRW) Application	https://github.com/ufs-community/ufs-srweather-app
Repository for the UFS Weather Model	https://github.com/ufs-community/ufs-weather-model
Repository for UFS Utilities, including pre-processing, chgres_cube, and more	https://github.com/ufs-community/UFS_UTILS
Repository for the Unified Post Processor (UPP)	https://github.com/NOAA-EMC/UPP
Repository for Air Quality Modeling (AQM) Utilities	https://github.com/NOAA-EMC/AQM-utils
Repository for the NOAA Emission and eXchange Unified System (NEXUS)	https://github.com/noaa-oar-arl/NEXUS
Repository for the Unified Workflow (UW) Toolkit	https://github.com/ufs-community/uwtools

The UFS Weather Model contains a number of sub-repositories, which are documented here.

Note

The prerequisite libraries (including NCEP Libraries and external libraries) are not included in the UFS SRW Application repository. The spack-stack repository assembles these prerequisite libraries. Spack-stack has already been built on preconfigured (Level 1) platforms. However, it must be built on other systems. See the spack-stack Documentation for details on installing spack-stack.

1.1.2.2.2. Repository Structure

The ufs-srweather-app umbrella repository is an NCO-compliant repository. Its structure follows the standards laid out in NCEP Central Operations (NCO) WCOSS Implementation Standards. This structure is implemented using the local_path settings contained within the Externals.cfg file. After manage_externals/checkout_externals is run (see Section 1.2.3.3), the specific GitHub repositories described in Table 1.2 are cloned into the target subdirectories shown below. Directories that will be created as part of the build process appear in parentheses and will not be visible until after the build is complete. Some directories have been removed for brevity.

ufs-srweather-app
├── (build)
├── docs
│     └── UsersGuide
├── etc
├── (exec)
├── (include)
├── jobs
├── (lib)
├── manage_externals
├── modulefiles
│     ├── build_<platform>_<compiler>.lua
│     └── wflow_<platform>.lua
├── parm
│     ├── wflow
│     │     └── default_workflow.yaml
│     └── FV3LAM_wflow.xml
├── (share)
├── scripts
├── sorc
│     ├── CMakeLists.txt
│     ├── (UPP)
│     │     ├── parm
│     │     └── sorc
│     │          └── ncep_post.fd
│     ├── (UFS_UTILS)
│     │     ├── sorc
│     │     │    ├── chgres_cube.fd
│     │     │    ├── sfc_climo_gen.fd
│     │     │    └── vcoord_gen.fd
│     │     └── ush
│     └── (ufs-weather-model)
│             └── FV3
│                ├── atmos_cubed_sphere
│                └── ccpp
├── tests/WE2E
│     └── run_WE2E_tests.py
├── ush
│     ├── bash_utils
│     ├── machine
│     ├── wrappers
│     ├── config.community.yaml
│     ├── config_defaults.yaml
│     ├── generate_FV3LAM_wflow.py
│     ├── launch_FV3LAM_wflow.sh
│     ├── setup.py
│     └── valid_param_vals.yaml
└── versions

1.1.2.2.2.1. SRW App SubDirectories

Table 1.3 describes the contents of the most important SRW App subdirectories. Table 1.8 provides a more comprehensive explanation of the ufs-srweather-app files and subdirectories. Users can reference the NCO Implementation Standards (p. 19) for additional details on repository structure in NCO-compliant repositories.

Table 1.3 *Subdirectories of the ufs-srweather-app repository*
Directory Name	Description
docs	Repository documentation
jobs	J-job scripts launched by Rocoto
modulefiles	Files used to load modules needed for building and running the workflow
parm	Parameter files used to configure the model, physics, workflow, and various SRW App components
scripts	Scripts launched by the J-jobs
sorc	External source code used to build the SRW App
tests	Tests for baseline experiment configurations
ush	Utility scripts used by the workflow

1.1.2.2.3. Experiment Directory Structure

When the user generates an experiment using the generate_FV3LAM_wflow.py script (Step 1.2.4.3.3), a user-defined experiment directory ($EXPTDIR) is created based on information specified in the config.yaml file. Table 1.4 shows the contents of the experiment directory before running the experiment workflow.

Table 1.4 *Files and subdirectory initially created in the experiment directory*
File Name	Description
config.yaml	User-specified configuration file, see Section 1.2.4.3.2.2
data_table	Cycle-independent input file (empty)
field_table	Tracers in the forecast model
FV3LAM_wflow.xml	Rocoto XML file to run the workflow
input.nml	Namelist for the UFS Weather Model
launch_FV3LAM_wflow.sh	Symlink to the `ufs-srweather-app/ush/launch_FV3LAM_wflow.sh` shell script, which can be used to (re)launch the Rocoto workflow. Each time this script is called, it appends information to a log file named `log.launch_FV3LAM_wflow`.
log.generate_FV3LAM_wflow	Log of the output from the experiment generation script (`generate_FV3LAM_wflow.py`)
nems.configure	See NEMS configuration file
suite_{CCPP}.xml	CCPP suite definition file (SDF) used by the forecast model
var_defns.sh	Shell script defining the experiment parameters. It contains all of the primary parameters specified in the default and user-specified configuration files plus many secondary parameters that are derived from the primary ones by the experiment generation script. This file is sourced by various other scripts in order to make all the experiment variables available to these scripts.
YYYYMMDDHH	Cycle directory (empty)

In addition, running the SRW App in community mode creates the fix_am and fix_lam directories (see Table 1.5) in $EXPTDIR. The fix_lam directory is initially empty but will contain some fix (time-independent) files after the grid, orography, and/or surface climatology generation tasks run.

Table 1.5 *Description of the fix directories*
Directory Name	Description
fix_am	Directory containing the global fix (time-independent) data files. The experiment generation script symlinks these files from a machine-dependent system directory.
fix_lam	Directory containing the regional fix (time-independent) data files that describe the regional grid, orography, and various surface climatology fields, as well as symlinks to pre-generated files.

Once the Rocoto workflow is launched, several files and directories are generated. A log file named log.launch_FV3LAM_wflow will be created (unless it already exists) in $EXPTDIR. The first several workflow tasks (i.e., make_grid, make_orog, make_sfc_climo, get_extrn_ics, and get_extrn_lbcs) are preprocessing tasks, and these tasks also result in the creation of new files and subdirectories, described in Table 1.6.

Table 1.6 *New directories and files created when the workflow is launched*
Directory/File Name	Description
YYYYMMDDHH	This is a “cycle directory” that is updated when the first cycle-specific workflow tasks (`get_extrn_ics` and `get_extrn_lbcs`) are run. These tasks are launched simultaneously for each cycle in the experiment. Cycle directories are created to contain cycle-specific files for each cycle that the experiment runs. If `DATE_FIRST_CYCL` and `DATE_LAST_CYCL` are different in the `config.yaml` file, more than one cycle directory will be created under the experiment directory.
grid	Directory generated by the `make_grid` task to store grid files for the experiment
log	Contains log files generated by the overall workflow and by its various tasks. View the files in this directory to determine why a task may have failed.
orog	Directory generated by the `make_orog` task containing the orography files for the experiment
sfc_climo	Directory generated by the `make_sfc_climo` task containing the surface climatology files for the experiment
FV3LAM_wflow.db FV3LAM_wflow_lock.db	Database files that are generated when Rocoto is called (by the launch script) to launch the workflow
log.launch_FV3LAM_wflow	The `launch_FV3LAM_wflow.sh` script appends its output to this log file each time it is called. View the last several lines of this file to check the status of the workflow.

The output files for an experiment are described in Section 1.3.2.2. The workflow tasks are described in Section 1.2.4.3.4.