Smoke & Dust Prediction Capability Announcement

[gspetro-NOAA]

Ravan Ahmadov^a, Haiqin Li^b, Johana Romero-Alvarez^b, Jordan Schnell^b, Sudheer Bhimireddy^b, Eric James^a, Ka Yee Wong^c, Ming Hu^a, Jacob Carley^d, Partha Bhattacharjee^d, Barry Baker^e, Georg Grell^a, Chuanyu Xu^f, Shobha Kondragunta^f, Fangjun Li^g, Samuel Trahan^b, Margaret Marvin^e, Chan-Hoo Jeon^h, Ben Koziol^h, Jong Kim^h, and Gillian Petro^h

^aNOAA/GSL, Boulder, USA; ^bCU Boulder, CIRES, Boulder, USA; ^cCSU/CIRA, Fort Collins, USA; ^dNOAA/EMC, College Park, USA; ^eNOAA/ARL, College Park, USA; ^fNOAA/NESDIS, College Park, USA; ^gSDSU, Brookings, USA; ^hNOAA/EPIC, College Park, USA

The Earth Prediction Innovation Center (EPIC) and the Unified Forecast System (UFS) community are proud to announce the addition of a smoke and dust prediction capability within the Short-Range Weather (SRW) Application. The Short-Range Weather – Smoke and Dust (SRW-SD) prediction capability is available on an SRW App feature branch: https://github.com/ufs-community/ufs-srweather-app/tree/main_aqm. SRW-SD is a port of the Rapid Refresh Forecast System – Smoke and Dust (RRFS-SD)^1,2,3 component and represents an important step in the SRW-RRFS convergence process. The main_aqm SRW App feature branch is targeted to be merged with SRW’s develop branch in 2025 as the UFS community works to resolve external RRFS dependencies.

Documentation for the new SRW-SD capability can be found in the SRW App User's Guide.

Smoke and dust modeling is an essential component of integrated modeling systems, as aerosol concentrations provide critical feedback in weather simulations and can adversely affect human health. In weather modeling, smoke and dust particles affect solar radiation, atmospheric temperatures, and precipitation patterns⁴. Smoke and dust particles also pose respiratory and cardiovascular risks, especially for vulnerable populations⁵. Releasing this new capability within the SRW App will accelerate scientific collaboration in the smoke and dust modeling community and pave the way for future improvements to the modeling system. This approach aligns with the UFS’s role as a community-driven Earth system modeling framework designed to advance weather forecasting by integrating research, development, and operational tools.

RRFS-SD (and SRW-SD) is a coupled Finite-Volume Cubed-Sphere Dynamical Core (FV3-coupled), offline aerosol model with three tracers of smoke, fine dust, and coarse particulate matter. Anthropogenic and biogenic emissions, gas, and aerosol chemistry are not included. RRFS-SD is the next generation of the HRRR-Smoke forecasting system⁶. The SRW-SD configuration can be simulated on a 3-km grid covering North America, and it produces deterministic and ensemble forecasts every hour out to 18 hours. RRFS-SD uses the High-Resolution Rapid Refresh (HRRR) as part of the FV3_HRRR_gf physics suite⁷. This physics suite is similar to the National Oceanic and Atmospheric Administration (NOAA) operational HRRR v4 suite^8,9, with the addition of the Grell-Freitas deep convective parameterization¹⁰. The SRW-SD capability produces the following fields:

• Hourly Wildfire Potential
• Smoke Emissions
• Near-Surface Smoke
• Vertically Integrated Smoke
• Near-Surface Fine Dust
• Near-Surface Course Dust
• Vertically Integrated Dust
• Visibility

Notable changes associated with the SRW-SD main_aqm branch include:

• Addition of three tasks to the SRW App workflow:
  • smoke_dust ⇒ Generates parameter files for smoke and dust for use by the UFS Weather Model.
  • prepstart ⇒ Adds smoke and dust fields to the initial conditions (ICs) files from the restart file in the previous cycle.
  • upp_post ⇒ Performs post-processing with the Unified Post Processor (UPP). This post-processor is NCO-compliant.
• Addition of diagnostic and prognostic field plotting capabilities related to aerosol and fire-related indices.
• Refactoring of the code structure to align with NCO standards.
• Addition of Python ush scripts to prepare smoke- and dust-related ICs for input into the UFS.

The SRW-SD capability underwent initial scientific validation. Interactive plots from a simulation spanning an August 1-10, 2019 forecast were created using the SRW-SD’s prognostic PM2.5 field and in situ measurements of PM2.5 aggregated by the MELODIES MONET¹¹ model evaluation framework. During this period, fires primarily impacted Washington State. The SRW-SD model only considers PM2.5 emissions from fires, not other sources such as vehicles. As a result, in areas without active fires, low PM2.5 values are output by the model.

Simulated hourly near-surface smoke and dust concentrations from an experimental SRW-SD forecast spanning August 1-10, 2019.

Future development pathways for the SRW-SD capability include full integration with the SRW App as part of a 2025 SRW App release, user-defined modeling domains, customized factor selection to tailor scientific experiments, surface data pre-processing flags for finer-scale controls of data quality, and full end-to-end testing against established baselines.

The main_aqm branch is tested on the following Tier-1 Research and Development High-Performance Computing (RDHPC) systems: Hera, Orion, and Hercules. Interested users can get support by submitting a question through the GitHub Discussions Q&A page.

Additional References

1. https://gsl.noaa.gov/research/experimental-models
2. https://macmaq.aqrc.ucdavis.edu/sites/g/files/dgvnsk6036/files/inline-files/3.3%20Ravan%20Ahmadov.pdf
3. https://github.com/ufs-community/ufs-weather-model/tree/production/RRFS.v1
4. Huang, Xin, and Aijun Ding. “Aerosol as a Critical Factor Causing Forecast Biases of Air Temperature in Global Numerical Weather Prediction Models.” Science Bulletin 66, no. 18 (September 30, 2021): 1917–24. https://doi.org/10.1016/j.scib.2021.05.009.
5. Liu, Yongqiang, Adam Kochanski, Kirk R. Baker, William Mell, Rodman Linn, Ronan Paugam, Jan Mandel, et al. “Fire Behavior and Smoke Modeling: Model Improvement and Measurement Needs for next-Generation Smoke Research and Forecasting Systems.” International Journal of Wildland Fire 28, no. 8 (2019): 570. https://doi.org/10.1071/wf18204.
6. https://rapidrefresh.noaa.gov/hrrr/HRRRsmoke/
7. https://dtcenter.ucar.edu/GMTB/v7.0.0/sci_doc/_h_r_r_r_gf_page.html
8. Dowell, David C., Curtis R. Alexander, Eric P. James, Stephen S. Weygandt, Stanley G. Benjamin, Geoffrey S. Manikin, Benjamin T. Blake, et al. “The High-Resolution Rapid Refresh (HRRR): An Hourly Updating Convection-Allowing Forecast Model. Part I: Motivation and System Description,” August 3, 2022. https://doi.org/10.1175/WAF-D-21-0151.1.
9. https://ccpp-techdoc.readthedocs.io/en/v7.0.0
10. Freitas, Saulo R., Georg A. Grell, Andrea Molod, Matthew A. Thompson, William M. Putman, Claudio M. Santos e Silva, and Enio P. Souza. “Assessing the Grell-Freitas Convection Parameterization in the NASA GEOS Modeling System.” Journal of Advances in Modeling Earth Systems 10, no. 6 (2018): 1266–89. https://doi.org/10.1029/2017MS001251.
11. https://melodies-monet.readthedocs.io/en/stable/