30 Sep (CLIM) - Tallapragada, NWP at NOAA NWS
Speaker: Dr. Vijay Tallapragada, NOAA NWS Office of Modeling and Development
Title:Operational Modeling Advancements in NOAA: Progress and Challenges
Time: Wed, 30 Sep, 1:30-2:30 pm
Location: Exploratory Hall, Room L111 or via Zoom (for Zoom link, email xdu5@gmu.edu)
Host: Jim Kinter
ABSTRACT: NOAA, in close collaboration with the broader community, is rapidly advancing a fully coupled Earth system modeling capability through the Unified Forecast System (UFS). As a community-based modeling framework, the UFS is designed to unify research and operations across a broad range of temporal and spatial scales. By replacing a collection of disparate legacy models with a unified modeling system, it supports applications spanning sub-hourly, convection-allowing local forecasts to subseasonal and seasonal global predictions, serving as the cornerstone of NOAA's strategy for next-generation environmental prediction.
In addition, Artificial Intelligence and Machine Learning (AI/ML) technology have advanced rapidly in recent years, transforming the field of numerical weather prediction with improved forecast skill and remarkable efficiency. In response, operational centers are beginning to integrate AI/ML models into their systems to provide transformative forecast products. Three distinct global AI applications (AIGFS, AIGEFS and HGEFS) have already been implemented into operations - these are NOAA's AI-based emulators of the Global Forecast System (GFS) and Global Ensemble Forecast System (GEFS), and HGEFS essentially combines physical and AI ensembles to generate a larger ensemble.
Thanks to the Administration’s renewed priority for global leadership in NWP, NOAA has embarked on a new Modeling Moonshot initiative (also known as modeling grand challenge). This is aimed at modernizing the operational modeling suite through a transformation to advanced software and cloud-based HPC architectures, expansion of hybrid data-driven (AI) & physics-based ensemble prediction systems, and the adoption of the Model for Prediction Across Scales (MPAS) for a global, coupled, km-scale ensemble system driven by a continuously-operating data assimilation (DA) system.
This presentation provides an update on several key UFS applications that are in advanced stages of development and are being prepared for operational implementation within the National Weather Service (NWS), along with AI based applications and the Moonshot Initiatives that are rapidly advancing.
