AI-Powered Climate Modeling Advances by the US NOAA
The US National Oceanic and Atmospheric Administration (NOAA) has made significant advances in AI-powered climate and weather modeling through initiatives such as Project EAGLE. Project EAGLE aims to provide NOAA and the broader US weather enterprise rapid capabilities to test, develop, and demonstrate AI-based models for global and regional ensemble forecasting in near-real time. This includes AI models tuned with NOAA data that operate on high-resolution grids to produce extended forecasts, enhancing both the speed and accuracy of weather predictions.
11/19/20252 min read
Artificial Intelligence Blog
What Is Project EAGLE?
Project EAGLE is NOAA’s experimental platform designed to develop, test, and deploy AI-based models for global and regional weather prediction. Leveraging vast NOAA datasets and cutting-edge AI techniques, EAGLE aims to complement and enhance traditional physics-based numerical models by harnessing the power of machine learning.
This innovative system includes two main forecast methods: deterministic forecasts that produce a single best prediction and ensemble forecasts that generate multiple possible scenarios to capture uncertainty and risk more effectively. By integrating these forecasts into NOAA’s existing suite of models, including the Global Forecast System (GFS) and the High-Resolution Rapid Refresh (HRRR), Project EAGLE accelerates the path from AI research to real-world weather prediction applications.
HRRR-Cast: NOAA’s First AI-Powered Regional Forecast
A pioneering achievement of Project EAGLE is HRRR-Cast, NOAA’s first AI-powered regional forecasting system. Trained on three years of data from NOAA’s physics-driven HRRR model, HRRR-Cast produces rapid, high-resolution forecasts focused on the continental U.S.
HRRR-Cast excels in predicting precipitation, temperature, wind, and storm structures, matching or exceeding the performance of the traditional HRRR in many respects. One of its major advantages is computational efficiency: HRRR-Cast runs significantly faster than conventional models, offering the potential for more frequent and timely updates, which is critical for severe weather warnings and emergency response.
Why AI Matters for Weather and Climate
Traditional weather models simulate atmospheric physics to forecast conditions, but these simulations demand extensive computing resources and time. AI models, however, learn from vast historical and real-time data to identify complex patterns and relationships woven throughout the atmosphere and oceans—leading to faster forecasts without sacrificing accuracy.
The probabilistic outputs from AI ensemble forecasts provide richer insights for decision-makers, emergency managers, and the public by clarifying forecast confidence and the range of possible outcomes. This fosters better preparation and response to extreme events such as hurricanes, flash floods, and heatwaves that are increasingly frequent under climate change.
Infrastructure and Collaborative Innovation
NOAA’s commitment to AI innovation goes hand-in-hand with investments in high-performance computing infrastructure and partnerships with industry leaders like Google DeepMind and Nvidia. This collaborative approach ensures NOAA remains at the forefront of AI applications in Earth system science.
Additionally, the Earth Prediction Innovation Center (EPIC) serves as an open research hub, enabling scientists globally to contribute to, test, and improve AI weather models. This joint effort enhances transparency, accelerates model improvements, and broadens community engagement.
Looking Ahead: Transforming Climate Modeling
Beyond short-term weather forecasting, NOAA is exploring how AI can enrich climate modeling. Machine learning helps refine critical model components, accelerate computations, and improve predictions of long-term climate patterns and ocean states. As these AI tools mature, they will be instrumental in addressing climate risks and supporting sustainability initiatives worldwide.
The advances NOAA has made with AI-powered weather and climate models represent a leap forward in forecasting science. With faster, accurate, and actionable predictions enabled by Project EAGLE and HRRR-Cast, NOAA is enhancing the nation’s resilience to weather hazards while setting a global example for integrating AI into environmental prediction.
Stay tuned as NOAA continues to push the boundaries of AI innovation in climate science, helping communities better understand and prepare for the challenges of a changing planet