12. Global Chemistry and Aerosol Forecast
12.1. Overview
The Global Chemistry and Aerosol Forecast System (GCAFS) extends the Global Forecast System (GFS) with interactive aerosol and atmospheric chemistry capabilities. It provides a unified framework for predicting the evolution of atmospheric composition alongside traditional weather variables.
12.2. Key Features
Interactive GOCART aerosol module for forecasting dust, sea salt, sulfate, black carbon, and organic carbon
Optional full atmospheric chemistry with gas-phase and heterogeneous reactions
Integration with biomass burning emissions sources (QFED, GBBEPX)
Aerosol-radiation-cloud interactions
Optional aerosol data assimilation
12.3. Running GCAFS
GCAFS can be run using the global-workflow framework. To set up a free-forecast GCAFS experiment:
./setup_expt.py gcafs forecast-only --pslot my_gcafs_run --app ATMA \
--idate 2023010100 --edate 2023010100 \
--resdetatmos 384 --comroot /path/to/com --expdir /path/to/exp
Configuration is managed through the standard global-workflow configuration files. GCAFS-specific settings are documented in gcafs_config.
After setting up the experiment, build the workflow XML and launch it:
./setup_workflow.py /path/to/exp/my_gcafs_run
cd /path/to/exp/my_gcafs_run
rocotorun -w gcafs.xml -d gcafs.db
One can also run in ‘cycled’ mode, see other sections of the global-workflow documentation for details on how to set up a cycled experiment. For GCAFS, the meteorological analysis in ‘cycled’ mode is taken from the GDAS (either from HPSS archive or a location stored on disk). The aerosol analysis is optional and controlled by setting USE_AERO_ANL to be “YES”
12.4. GCAFS Workflow
The GCAFS workflow includes these main tasks:
stage_ic - Stage initial conditions
prep_emissions - Prepare emissions data files
aerosol_init - Initialize aerosol fields
fcst - Run the UFS model with aerosols/chemistry
atmos_prod - Post-process atmosphere/aerosol output
arch_vrfy and arch_tars - Archive verification data and create tarballs
The workflow is managed by the Rocoto workflow manager, with tasks defined in the
workflow/rocoto/gcafs_tasks.py
file.
12.5. Emissions Preprocessing
The prep_emissions
task is a critical component of the GCAFS workflow that prepares all necessary emissions data and configuration files for the model run.
### Overview of prep_emissions
This task performs several important functions:
Configuration Generation: Creates customized GOCART configuration files from templates
Emissions File Preparation: Processes and prepares emissions data files
Historical Data Handling: Retrieves historical fire emissions when needed
Fire Emissions Selection: Configures the selected biomass burning emissions source (QFED/GBBEPx)
Template Variable Processing: Processes all template variables in the configuration files
The task is implemented in ush/python/pygfs/task/aero_emissions.py
as the AerosolEmissions
class.
### Detailed Workflow
When the prep_emissions
task runs, it follows these steps:
Initialization: ```python def initialize(self):
# Parse the YAML template for chemistry emissions yaml_template = os.path.join(self.task_config.HOMEgfs, ‘parm/chem/chem_emission.yaml.j2’) yamlvars = parse_j2yaml(path=yaml_template) self.task_config.append(yamlvars)
This loads the base configuration template and merges it with the task configuration.
Historical Fire Emission Handling: ```python if self.task_config.fire_emissions == ‘historical’:
# Handle historical fire emissions self.task_config.fire_emissions = ‘historical’ self.task_config.fire_emissions_file = os.path.join(self.task_config.HOMEgfs, ‘parm/chem/historical_fire_emissions.txt’)
This sets up the task to use historical fire emissions data if specified.
Fire Emission Configuration: ```python if self.task_config.fire_emissions == ‘qfed’:
# Configure QFED emissions self.task_config.fire_emissions = ‘qfed’ self.task_config.fire_emissions_file = os.path.join(self.task_config.HOMEgfs, ‘parm/chem/qfed_fire_emissions.txt’)
This sets up the task to use QFED emissions data if specified.
12.6. GOCART Configuration Files
The GOCART aerosol module in GCAFS is configured through a set of resource (.rc) files located in
parm/ufs/gocart/
. These files control the behavior of the aerosol components, emissions,
and diagnostics. The key configuration files include:
12.6.1. Core Configuration
AERO.rc: Core aerosol module configuration with grid resolution settings
AGCM.rc: Atmospheric model interface configuration
CAP.rc: Component interface specifications defining imports/exports and tracer mappings
GOCART2G_GridComp.rc: Defines active aerosol species instances (DU, SS, SU, CA, NI)
12.6.2. Aerosol Species Configuration
Each aerosol species has its own configuration file with specific parameters:
DU2G_instance_DU.rc: Dust particle properties, emission scheme settings (fengsha/ginoux/k14)
SS2G_instance_SS.rc: Sea salt configuration with size bins and emission methods
SU2G_instance_SU.rc: Sulfate species configuration including volcanic emissions
CA2G_instance_CA.bc.rc: Black carbon aerosol properties and behavior
CA2G_instance_CA.oc.rc: Organic carbon configuration including SOA formation
NI2G_instance_NI.rc: Nitrate aerosol specification (optional species)
12.6.3. Output and Diagnostics
AERO_HISTORY.rc: Controls aerosol output diagnostics including: - Aerosol concentrations (inst_du_ss, inst_ca, inst_ni, inst_su) - Process-specific outputs (emission, deposition) - Optical properties (AOD calculations) - Output grid configuration and file formats
The frequency parameters for output are specified as variables (e.g., @[inst_aod_freq]
) that
are replaced at runtime with values from the workflow configuration.
12.6.4. Emissions Configuration
External data sources for emissions are configured in:
ExtData.gbbepx: GBBEPx biomass burning emissions configuration
ExtData.qfed: QFED fire emissions configuration
ExtData.other: Anthropogenic, biogenic, and other emission sources
ExtData.none: Placeholder configuration when emissions are disabled
To modify the aerosol configuration, edit these files or create custom versions in your experiment
directory. The file gocart_tracer.list
defines the complete set of aerosol tracers used in the model.
12.6.5. ExtData File Format Details
The ExtData configuration files specify how external data sources are imported into the model. Each entry follows this format:
# Import Name | Units | Clim | Regrid | Time Template | Offset | Scale | Var on File | File Template
OC_BIOMASS NA N Y %y4-%m2-%d2t12:00:00 none 0.7778 biomass ExtData/nexus/QFED/%y4/%m2/qfed2.emis_oc.006.%y4%m2%d2.nc4
Field descriptions:
Import Name: Internal variable name used by GOCART (e.g., OC_BIOMASS, SU_BIOMASS)
Units: Units for the imported field or NA if not applicable
Clim: Climate file flag (Y/N) - Y indicates a static/climatological file
Regrid: Regrid method - Y for bilinear interpolation, N for no regridding, E for conservative regridding
Time Template: Date/time template for file selection using tokens (%y4, %m2, %d2 for year, month, day)
Offset Factor: Value to add to data after reading (or “none”)
Scale Factor: Value to multiply data by after reading (or “none”)
Variable On File: Name of the variable in the source file
File Template: Path to source file with date tokens for time-varying data
Time tokens include: - %y4 - 4-digit year - %m2 - 2-digit month - %d2 - 2-digit day - %h2 - 2-digit hour - %n2 - 2-digit minute
For example, in the QFED configuration:
SU_BIOMASS NA N Y %y4-%m2-%d2t12:00:00 none 0.7778 biomass ExtData/nexus/QFED/%y4/%m2/qfed2.emis_so2.006.%y4%m2%d2.nc4
This imports SO2 emissions from QFED into the SU_BIOMASS variable, using a scale factor of 0.7778, from files with a date-based naming pattern.
12.6.6. AERO_HISTORY.rc File Details
The AERO_HISTORY.rc file controls all diagnostic outputs from the aerosol module. It defines:
Collections: Groups of variables output together in a single file
Grid configuration: Output grid specifications (resolution, type)
Output frequency: How often each collection is written
File formats and templates: How output files are named and formatted
Basic structure:
COLLECTIONS: 'inst_aod'
'inst_du_ss'
'inst_ca'
...
GRID_LABELS: PC720x361-DC
::
PC720x361-DC.GRID_TYPE: LatLon
PC720x361-DC.IM_WORLD: 720
PC720x361-DC.JM_WORLD: 361
PC720x361-DC.POLE: PC
PC720x361-DC.DATELINE: DC
For each collection, several parameters are defined:
inst_aod.format: 'CFIO'
inst_aod.template: '%y4%m2%d2_%h2%n2z.nc4'
inst_aod.mode: 'instantaneous'
inst_aod.grid_label: PC720x361-DC
inst_aod.frequency: @[inst_aod_freq]
inst_aod.duration: 010000
inst_aod.fields: 'DUEXTTAU' , 'GOCART2G' ,
'SSEXTTAU' , 'GOCART2G' ,
...
Where:
format: Output file format (‘CFIO’ for NetCDF)
template: File naming template with time tokens
mode: Output type (‘instantaneous’ or ‘time-averaged’)
grid_label: Reference to a grid defined in GRID_LABELS
frequency: How often to output (specified as a variable like @[inst_aod_freq])
duration: How long this collection is active (HHMMSS format)
fields: List of variable name pairs (internal name, component name)
The frequency parameters use variables like @[inst_aod_freq]
that are replaced at runtime with
values from the workflow configuration. These variables are typically set in the emissions
preprocessing step and use the format HHMMSS (hours, minutes, seconds).
Common collections include:
inst_aod: Instantaneous aerosol optical depth
inst_du_ss: Dust and sea salt concentrations
inst_ca: Carbonaceous aerosol (BC/OC) concentrations
inst_su: Sulfate species concentrations
inst_3d: 3D fields for all aerosols
inst_2d: 2D diagnostic fields (column mass, surface concentrations)
To enable specific collections, uncomment them in the COLLECTIONS section and ensure the corresponding frequency parameters are properly set in your workflow.
12.7. Output Products
GCAFS produces standard meteorological outputs plus aerosol fields including:
Aerosol mass concentrations (dust, sea salt, sulfate, black carbon, organic carbon)
Aerosol optical depth fields
PM2.5 and PM10 concentrations
Full chemical species concentrations when running with chemistry enabled
Output frequency is controlled by the standard global-workflow configuration options in the same manner as GFS.