the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 20 Apr 2026
Development of the CCPP-Based GEFS-Aerosols Component in the Unified Forecast System for Subseasonal Prediction (UFS-Chem v1.0)
Abstract. The Global Ensemble Forecast System (GEFS) version 12 has been operational at the National Centers for Environmental Prediction since September 2020, with GEFS-Aerosols serving as its global aerosol forecasting member. In 2023, GEFS-Aerosols was upgraded to version 12.3 in operations, incorporating improvements to wet deposition, anthropogenic and biomass burning emissions, aerosol optical depth calculations, and the FENGSHA dust emission scheme. While GEFS-Aerosols provides valuable operational aerosol forecasts on medium-range timescales, its one-way coupling strategy restricts aerosol–atmosphere interactions to prescribed climatological fields, thereby precluding fully interactive aerosol feedbacks for extended-range prediction. To overcome this limitation, we develop the Unified Forecast System coupled with Chemistry (UFS-Chem), which incorporates the Configurable Atmospheric Chemistry (CATChem) library and modeling component to include both aerosol and gas-phase chemistry schemes. In the current UFS-Chem configuration, the aerosol component is based on the operational GEFS-Aerosols v12.3 and implemented using the Common Community Physics Package (CCPP) framework. Relative to GEFS-Aerosols v12.3, UFS-Chem incorporates several enhancements, including inline aerosol radiative feedback, inline large-scale wet deposition, and updated FENGSHA dust scheme. In particular, the integration of inline large-scale wet deposition and aerosol indirect effects through the Thompson aerosol-aware microphysics scheme improves the representation of aerosol–cloud interactions and supports weather and subseasonal-to-seasonal forecasting when fully coupled with ocean, sea ice, land, and wave components. The performance of UFS-Chem in weather and subseasonal prediction is evaluated against reanalysis data, ground-based measurements, and satellite observations. The relative impacts of aerosol radiative and indirect feedback on weather and subseasonal predictions are investigated and quantified, advancing our understanding of how aerosol–radiation and aerosol–cloud interactions influence forecast skill across timescales.
- Preprint
(6795 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 15 Jun 2026)
- RC1: 'Comment on egusphere-2026-1807', Anonymous Referee #1, 19 May 2026 reply
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 238 | 108 | 21 | 367 | 17 | 16 |
- HTML: 238
- PDF: 108
- XML: 21
- Total: 367
- BibTeX: 17
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This paper presents UFS-Chem, a new forecast model with inline aerosol–radiation and aerosol–cloud interactions under the CCPP framework. Overall, the manuscript is well written and clearly structured. However, I have several major concerns that require clarification prior to publication.
Major comments:
1. What is the computational cost for UFS-Chem compared with v12.3? It would be useful to provide a quantitative number
2. Why are all the discussion periods in August only? The balance between direct and indirect aerosol effects varies by season.
3. Why do Figures 2–3 present the time period of August 2021, but Figures 4 and 6 are for August 2016? The inconsistency in evaluation periods is confusing. Also, why do Figures 2–3 focus on v12.3 rather than demonstrating UFS-Chem performance directly?
4. Figure 4: I wonder why CYC_IWR generates such a huge bias near central Africa? Do you have any explanation for this?
5. The transition from v12.3 to UFS-Chem incorporates changes to the coupling framework, physics suite (GFSv15→GFSv17), vertical levels (65→127), microphysics, land model, and convection scheme simultaneously. The paper cannot attribute benefits to CCPP coupling specifically versus the physics upgrade. Please acknowledge this limitation explicitly.
6. Figures 10–12: Why does EXP1 improve the cloud fraction bias but worsen the TOA flux and SST biases, whereas EXP2 corrects all of them? Do the EXP2 improvements reflect better physics or a new balance of compensating errors?
7. Figure 11: A better evaluation for the indirect effect implementation would be to compare against cloud liquid water path and droplet effective radius.
Minor comments:
Figure 6: please add quantitative annotations (global mean bias) on each panel. It is hard to tell which experiment performs better as presented.
Table 4: It would be more effective to present Table 4 as a scatter plot. Also, do you have any explanation for the few stations where v12.3 degrades?
L288: add a reference for the Baker–Schepanski map
L288–289: What is the sea salt emission scheme? Does wave model coupling influence sea salt production?
L306–307: Wave model uncoupling for the 6-year experiments: does this affect SST or surface flux results?
L376: What are the mass-to-number conversion parameters for mapping GOCART mass to CCN/INP?
Typos: Table 1: First column, sixth row: “GEFS-COCART” -> “GEFS-GOCART"; “CDES”-> “CEDS”