Abstract

Over the last 18 months, a tidal wave of data-driven models revolutionized the way we think about the future of the weather prediction enterprise (Bauer et.al 2023). The potential impact of machine learning (ML) on our ability to forecast weather has been compared to the advent of numerical weather prediction itself. Specifically, groundbreaking data-driven models (Pathak et.al 2022, Bi et.al 2023, Lam et.al 2023) showed that ML emulators trained on the ERA5 reanalysis are capable of achieving the skill of the flagship (high-resolution) deterministic weather forecast from the ECMWF at ~1/1,000 cost of the traditional model. Early results also show enhanced skill of the new data-driven models in predicting extreme life threatening events like hurricanes, winter storms, and heat waves.

NOAA's mission is to understand how our climate, weather, and oceans are changing, issue effective forecasts of these changes from minutes to decades into the future, and disseminate this information to the Nation to promote commerce, save life, and save livelihoods. For NOAA to effectively deliver on its mission, it is imperative for NOAA to embrace the emerging data-driven modeling technology to deliver environmental forecasts and datasets to the Nation more rapidly, at a lower cost, and with more skill. The velocity with which the field of data-driven models is developing is unprecedented and urges NOAA to re-imagine how innovations are incorporated into operational practice and how NOAA delivers the forecast products to the public. Without embracing the new technology, NOAA will fall behind the peer weather forecasting centers and the private sector, which will leave the Nation’s people and commerce in a competitive disadvantage and under threat from changing climate and extreme weather conditions.


This flash seminar will summarise findings of the workshop conducted by NOAA PSL, GSL, OAR, and EMC in November of 2023 to find pathways and opportunities for bringing data driven models within the NOAA research to operation pipeline for numerical weather prediction.