A broad collaboration among NOAA and non-NOAA scientists was launched in July 2020 to accelerate innovation into NOAA operational modeling for weather and climate prediction. The Unified Forecast System (UFS), a community-based, coupled comprehensive end-to-end Earth system data assimilation and prediction system, is being used in a Research-to-Operations (R2O) project. The UFS-R2O project was established in response to advice from the community that NOAA modeling and data assimilation needs to be integrated and collectively managed, based on a unified modeling framework in a unified collaborative strategy. The project, an experiment to carry out research and development collaboratively across the community within the constraints imposed by operational imperatives and public release timelines, is being supported by coordinated two-year funding from both NOAA National Weather Service’s Office of Science and Technology Integration and NOAA Oceanic and Atmospheric Research’s Weather Program Office. While the UFS-R2O Project has a two-year lifetime, the work on the 3-5 year vision starts immediately so that the pipeline of innovations that can be transitioned to operations is continuously fed. The project priorities are drawn from both forecaster requirements and the latest scientific developments.

Check out the new UFS R2O Project website  for more details.  

This ambitious project will be developing two major forecast systems, a coupled global ensemble system suitable for medium to extended range forecast (days to weeks) and an ensemble regional forecast system suitable for short-range forecast (hours to days). Both major systems are targeting operational implementation in 2024 and will be the first operational systems to be developed by an integrated research/operational and NOAA/non-NOAA community. Public releases of the two systems will also be suitable for use by the research community.

Further detail on the project

The UFS R2O Project includes two major deliverables:
First, this project aims to produce a fully coupled global ensemble Earth system prediction system – with active components representing the atmosphere, land surface, ocean, including ocean surface waves, sea ice, and aerosols. There will be public releases of this system for medium-range weather forecasting and for subseasonal-to-seasonal prediction. There will also be a public release of the Joint Effort for Data Assimilation Integration (JEDI) data assimilation framework and a reanalysis and reforecast capability for calibration and bias correction as well as state estimation. The second major outcome will be a regional rapid refresh ensemble forecast system (RRFS) that resolves convective-allowing scales. There will be a public release of RRFS for collaborative research.

Beyond the two-year horizon, the UFS-R2O Project will lay the groundwork for a strongly-coupled data assimilation capability for global Earth system prediction, a unified atmospheric physics suite for both convective-allowing and global modeling applications, an inline air quality prediction system at convective-allowing scales for the U.S., including aerosol feedback on global prediction applications, a Warn-on-Forecast system for use in predicting hazardous weather and significant flash flooding events, a Hurricane Analysis & Forecast System (HAFS) with moving nests following multiple storms, and a space-weather application. All innovations will be documented in research publications in high-impact peer-reviewed journals.

There are over 200 participants in the UFS-R2O Project from inside and outside NOAA under a single management framework with three co-equal principal investigators: Jeff Whitaker (NOAA ESRL PSD), Vijay Tallapragada (NOAA NCEP EMC) and Jim Kinter (George Mason University). The project is organized into two application teams focusing on (1) global modeling for medium-range weather forecasting and subseasonal to seasonal prediction, and (2) convective-allowing modeling for short-range weather forecasting and hurricanes. A third team is developing the cross-cutting software and data infrastructure to support the application teams in both modeling and analysis. Each team has several sub-projects focusing on atmospheric model physics and composition, data assimilation, coupled modeling, rapid refresh forecasting, and hurricane analysis and forecasting. The work in the UFS-R2O Project will be integrated into the broader UFS, and will leverage projects funded by NOAA NOFOs, NOAA research and development high-performance computing, and the NOAA cloud computing initiative.

The technical approach for the UFS-R2O Project involves coordinated development across NOAA and university partners of shared modelling and data assimilation infrastructure and algorithms, including frameworks for data assimilation (JEDI), coupling model components (ESMF, NUOPC, CMEPS), interoperable atmospheric sub-grid physical parameterizations (CCPP), community data models (CDEPS), and verification and validation framework (METPlus). All software development is conducted under UFS community code management policies, including managing Git-based repositories with Gitflow. A community workflow (CIME Case Control) is being extended and utilized. There is extensive attention to documentation and user support in coordination with the UFS Communications & Outreach working group. The UFS-R2O Project will coordinate with operations at EnvironModeling Center operations to streamline and accelerate the transition to operations, initially targeting Global Forecast System (GFS) v17, Global Ensemble Forecast System (GEFS) v13, and High Resolution Ensemble Forecast (HREF) v3.

A public-version of the project proposal is available here.

Project Leads: Jim Kinter (ikinter@gmu.edu), Vijay Tallapragada (vijay.tallapragada@noaa.gov) and Jeffrey Whitaker (jeffrey.whitaker@noaa.gov)
NOAA Program office: Dorothy Koch (dorothy.koch@noaa.gov, NWS-OSTI), DaNa Carlis (dana.carlis@noaa.gov, OAR-WPO) and Chandra Kondragunta (chandra.kondragunta@noaa.gov, OAR-WPO)