GCAFS Overview - TerrenceMcGuinness-NOAA/global-workflow GitHub Wiki

GCAFS (Global Composition/Chemistry Aerosol Forecast System)

Report Date: January 30, 2026
Repository: https://github.com/NOAA-EMC/GCAFS
Analysis Method: MCP/RAG tools + Git history analysis

Executive Summary

GCAFS (Global Composition/Chemistry Aerosol Forecast System) is NOAA's next-generation aerosol and air quality forecasting capability, built on the proven global-workflow infrastructure. It represents the fourth major forecasting system alongside GFS, GEFS, and SFS.

Aspect	Details
Purpose	Aerosol/air quality forecasting with GOCART model integration
Status	Active development, feature branch merged March 2025
Repository	Fork of global-workflow with GCAFS-specific configurations
Lead Developers	Barry Baker, Li Pan, Cory Martin
Total Commits	4,040 (inherited from global-workflow history)
First GCAFS Commit	March 12, 2025

What is GCAFS?

GCAFS is a specialized application of the global-workflow framework focused on aerosol transport, chemistry, and air quality prediction. While GFS predicts weather and GEFS provides ensemble forecasts, GCAFS focuses on:

Atmospheric aerosol concentrations (dust, smoke, sea salt, sulfate, organic/black carbon)
Air quality forecasting for public health advisories
Aerosol-radiation interactions affecting weather prediction
Fire emissions impacts on atmospheric composition

Key Distinctions from Other Systems

System	Purpose	Data Assimilation	Key Model
GFS	Medium-range weather	Full atmospheric DA (GSI/JEDI)	FV3
GEFS	Ensemble weather prediction	Ensemble DA	FV3
SFS	Seasonal/subseasonal forecasting	Seasonal cycles	FV3
GCAFS	Aerosol/air quality forecasting	Aerosol-specific DA	FV3 + GOCART

Repository Architecture

Relationship to global-workflow

GCAFS is a fork of NOAA-EMC/global-workflow that maintains parallel development:

NOAA-EMC/global-workflow (upstream)
        │
        ├──── NOAA-EMC/GCAFS (GCAFS-specific fork)
        │
        └──── Both repos actively syncing GCAFS features

Evidence of bidirectional development:

From global-workflow develop:

3db7edff "Enable gcafs tests on Ursa (#4439)"
5091a7b6 "Adding GCAFS products v2 (#4407)"
b446d9fd "Process aerosol analysis in grib2 format via UPP in GCAFS (#4234)"

From GCAFS repo:

136322b5 "Initial GCAFS configuration and application"
52b37f3e "Add GCAFS support in configuration and workflow files"
d0824cbe "Update GOCART path for GDAS/GFS/GCAFS implementations"

Submodule Dependencies

GCAFS inherits all global-workflow submodules:

Submodule	Repository	Purpose
`ufs_model.fd`	ufs-community/ufs-weather-model	UFS Weather Model core
`gdas.cd`	NOAA-EMC/GDASApp	JEDI-based data assimilation
`gsi_enkf.fd`	NOAA-EMC/GSI	GSI/EnKF data assimilation
`wxflow`	NOAA-EMC/wxflow	Workflow utilities
`gfs_utils.fd`	NOAA-EMC/gfs-utils	GFS-specific utilities
`ufs_utils.fd`	ufs-community/UFS_UTILS	UFS preprocessing utilities
`gsi_utils.fd`	NOAA-EMC/GSI-Utils	GSI utilities
`gsi_monitor.fd`	NOAA-EMC/GSI-Monitor	GSI monitoring tools
`verif-global.fd`	NOAA-EMC/EMC_verif-global	Verification tools

GCAFS-Specific Configuration

The initial GCAFS commit (136322b5) introduced 70+ configuration files in parm/config/gcafs/:

Configuration Categories

Category	Files	Purpose
Aerosol Analysis	`config.aero`, `config.aeroanl`	Aerosol DA configuration
Atmospheric Analysis	`config.atmanl`, `config.atmensanl`	ATM analysis with aerosol integration
Forecast	`config.fcst`, `config.efcs`	Forecast model settings
Products	`config.atmos_products`, `config.awips`	Output product generation
Archiving	`config.arch_`, `config.earc_`	HPSS/Globus archiving
Base/Common	`config.base`, `config.com`	Core GCAFS settings

Key GCAFS Base Configuration

From parm/config/gcafs/config.base:

#! /usr/bin/env bash
## THIS IS GCAFS
########## config.base ##########

# Toggle for GCAFS-specific features
export DO_GOES="@DO_GOES@"               # GOES products
export DO_BUFRSND="@DO_BUFRSND@"         # BUFR sounding products
export DO_GEMPAK="@DO_GEMPAK@"           # GEMPAK products
export DO_AWIPS="@DO_AWIPS@"             # AWIPS products
export DO_TRACKER="@DO_TRACKER@"         # Hurricane track verification
export DO_GENESIS="@DO_GENESIS@"         # Cyclone genesis verification

Development Evolution Timeline

Phase 1: FV3GFS Origins (2016-2017)

The repository traces back to the original FV3GFS project:

2016-10-18  d1d05ffd  "Creating new project: fv3gfs Trunk"
2016-10-19  019054cb  "FV3GFS runs on Cray"
2017-01     124 commits - Heavy initial development

Key milestone: Transition from spectral GFS to FV3-based GFS with cubed-sphere grid.

Phase 2: GFS v16 Development (2018-2020)

2018-09     178 commits - Peak activity, major infrastructure work
2019-2020   Steady development toward GFS v16 operational

Key milestone: GFS v16 operational implementation.

Phase 3: JEDI Integration & Multi-System Support (2021-2024)

2022-07-22  ffcd5bbd  "Add GDASapp (first wave of JEDI changes) (#871)"
2023-01-20  9fc82753  "Add initial Sphinx documentation (#1258)"
2023-02-10  1040216d  "Add in initial 3DVar aerosol DA cycling capability (#1106)"
2024-01-11  a65e4c67  "Initial update of version/module files #2123"

Key milestone: JEDI-based data assimilation begins replacing legacy GSI for some applications.

Phase 4: GCAFS Introduction (March 2025)

2025-03-12  136322b5  Barry Baker: "Initial GCAFS configuration and application"
2025-03-12  d8d9f359  Li Pan: "modification of the global-workflow for GCAFS"
2025-03-13  67c6dc56  Barry Baker: "edit parm/configs/gcafs"
2025-03-14  52b37f3e  Barry Baker: "Add GCAFS support in configuration and workflow files"
2025-03-18  d45b886d  Barry Baker: "minor updates for gcafs configuration"
2025-03-21  d0824cbe  Barry Baker: "Update GOCART path for GDAS/GFS/GCAFS implementations"
2025-04-09  4e9959da  "Add initial capability to produce JEDI-based observation space summary stat files"

Key milestone: GCAFS becomes a first-class citizen alongside GFS, GEFS, and SFS.

Git Statistics

Commit Volume by Year

Year	Commits	Notable Events
2016	37	Project creation
2017	430	Initial development surge
2018	509	GFS v16 preparation
2019	291	Stabilization
2020	269	Operations preparation
2021	261	Post-v16 development
2022	349	JEDI integration begins
2023	383	Multi-system expansion
2024	407	SFS and GCAFS development
2025 (Q1)	160+	GCAFS introduction

Active Branches

Branch	Purpose
`develop`	Main development (default)
`feature/gcafs_enable`	GCAFS feature development
`gcafsv1`	GCAFS version 1 release
`pantemp`	Li Pan's development branch

Top Contributors (All Time)

Rank	Developer	Commits
1	Fanglin Yang	775
2	Kate Friedman	639
3	Russ Treadon	233
4	Rahul Mahajan	196
5	Walter Kolczynski	171
6	George Gayno	174
7	David Huber	132
8	Mallory Row	113
9	Terrence McGuinness	97
10	Boi Vuong	83

GCAFS-Specific Contributors

Developer	Role
Barry Baker (`bbakernoaa`)	Lead GCAFS architect
Li Pan	GCAFS modifications
Cory Martin	Aerosol DA cycling
lipan-NOAA	Aerosol radiative feedback

Supported Platforms

Platform	Tier	Support Level
WCOSS2 (NCO)	1	Full EMC support, production target
Hera (NOAA RDHPCS)	1	Full EMC support, primary dev/test
Hercules (MSU)	1	Full EMC support (no TC Tracker)
Orion (MSU)	2	Ad-hoc support (GSI runs slowly)
Gaea C5/C6 (RDHPCS)	3	EPIC-supported
AWS/GCP/Azure (PW)	3	EPIC-supported cloud

Tier Definitions:

Tier 1: Full EMC support with regular CI testing
Tier 2: Ad-hoc EMC support, CI supported but not regular
Tier 3: Community/EPIC supported, no EMC guarantee

Aerosol Data Assimilation Evolution

The aerosol DA capability that GCAFS leverages has been building for years:

2022-08-19  8a62c3a9  "Add initial config files for global aerosol DA (#986)"
2023-02-10  1040216d  "Add in initial 3DVar aerosol DA cycling capability (#1106)"
2023-05-01  6c48e94b  "Update aerosol DA to use new COM structure (#1551)"
2023-07-27  de7b10aa  "Make small changes to get (aerosol) cycling going at C384"
2023-08-31  573ecce8  "Changes to get C48 aerosol test ready (#1829)"
2023-12-07  3483f550  "Add support for FGAT and IAU for aerosol DA (#1940)"
2024-06-14  6c933455  "Add observation preparation job for aerosols DA (#2624)"
2024-08-29  7f160f66  "Add diffusion/diag B for aerosol DA (#2738)"
2024-09-06  d3ea8e29  "Update aerosol climatology to 2013-2024 mean (#2888)"
2024-11-22  313a461e  "Make aerosol output frequency variable (#2982)"
2024-12-13  bef3e0ba  "Switch Aerosol DA to use JCB and Jedi class (#3125)"
2025-02-14  0da38327  "Fix the empty aerosol DA aerostat tar file issue (#3332)"
2025-02-21  586f658d  "Allow for C1152 ATM-Aero cycled DA to run on WCOSS2 (#3309)"

Job Scripts

GCAFS leverages the extensive job script infrastructure:

Atmosphere Jobs

JGDAS_ATMOS_ANALYSIS - GDAS atmospheric analysis
JGLOBAL_ATM_ANALYSIS_* - Global atmospheric analysis suite
JGLOBAL_ATMENS_ANALYSIS_* - Ensemble atmospheric analysis

Aerosol-Specific Jobs

JGLOBAL_AERO_ANALYSIS_INITIALIZE - Aerosol analysis initialization
JGLOBAL_AERO_ANALYSIS_VARIATIONAL - 3DVar aerosol analysis
JGLOBAL_AERO_ANALYSIS_FINALIZE - Aerosol analysis finalization
JGDAS_AERO_ANALYSIS_GENERATE_BMATRIX - Background error covariance

Product Jobs

JGFS_ATMOS_GEMPAK* - GEMPAK products
JGFS_ATMOS_AWIPS* - AWIPS products
JGFS_ATMOS_POSTSND - Post-processing sounding

Integration with GOCART

GCAFS integrates with the GOCART (Goddard Chemistry Aerosol Radiation and Transport) model for aerosol physics:

2025-03-21  d0824cbe  "Update GOCART path for GDAS/GFS/GCAFS implementations (#3455)"

GOCART provides:

Dust emission and transport
Sea salt aerosol generation
Sulfate chemistry
Carbonaceous aerosol (black carbon, organic carbon)
Biomass burning emissions
Aerosol optical depth calculations

Operational Readiness Status

Current State (January 2026)

Capability	Status
Configuration files	✅ Complete (70+ files)
Workflow integration	✅ Merged to develop
CI testing	🚧 Being enabled (Ursa, MSU)
Products v2	✅ Added
GRIB2 processing	✅ Working via UPP
WCOSS2 support	✅ Validated
C1152 resolution	✅ Supported

Upcoming Milestones

Based on git activity and PR patterns:

Q1 2026: Complete CI test coverage on all Tier 1 platforms
Q2 2026: Pre-operational testing on WCOSS2
Q3/Q4 2026: Potential operational implementation

Key Takeaways

GCAFS is not new code - It's a configuration of the mature global-workflow infrastructure for aerosol forecasting
Parallel development - Both NOAA-EMC/GCAFS and NOAA-EMC/global-workflow repos actively track GCAFS features
JEDI integration - GCAFS leverages the modern JEDI data assimilation framework for aerosol analysis
Production pathway - Following the same CI/CD and operational readiness process as GFS v17
Multi-platform support - Inherits all global-workflow platform support (WCOSS2, Hera, Hercules, cloud)

References

GCAFS Repository: https://github.com/NOAA-EMC/GCAFS
global-workflow Repository: https://github.com/NOAA-EMC/global-workflow
Documentation: https://global-workflow.readthedocs.io/
GOCART Model: https://github.com/GEOS-ESM/GOCART

Report generated using MCP/RAG analysis tools from the EIB MCP Gateway server
Analysis performed: January 30, 2026