Washington/CMEDIA: A groundbreaking new suite of operational, artificial intelligence (AI)-driven global weather prediction models has reportedlt been launched by the National Oceanic and Atmospheric Administration (NOAA) marking a significant advancement in forecast speed, efficiency, and accuracy.
While using only a fraction of computational resources, the models will provide forecasters with faster delivery of more accurate guidance.
“NOAA’s strategic application of AI is a significant leap forward in American weather model innovation,” said Neil Jacobs, Ph.D., NOAA administrator. “These AI models reflect a new paradigm for NOAA in providing improved accuracy for large-scale weather and tropical tracks, and faster delivery of forecast products to meteorologists and the public at a lower cost through drastically reduced computational expenses.”
Three distinct applications are included in the new suite of AI weather models:
- AIGFS (Artificial Intelligence Global Forecast System): Implementing AI to deliver improved weather forecasts, this weather forecast model more quickly and efficiently (using up to 99.7% less computing resources) than its traditional counterpart.
- AIGEFS (Artificial Intelligence Global Ensemble Forecast System): This ensemble system based on AI provides a range of probable forecast outcomes to meteorologists and decision-makers with improved performance, early results show, over the traditional GEFS, extending forecast skill by an additional 18 to 24 hours.
- HGEFS (Hybrid-GEFS): Combines the new AI-based AIGEFS (above) with NOAA’s flagship ensemble model, the Global Ensemble Forecast System, pioneering, hybrid “grand ensemble” that this model, a first-of-its kind approach for an operational weather center, consistently outperforms both the AI-only and physics-only ensemble systems.
More about the new AI operational models
AIGFS: A variety of data sources used by AIGFS to generate weather forecasts to generate weather forecasts comparable to those produced by traditional weather prediction systems, such as GFS.
- Performance: Improved forecast skill shown over the traditional GFS for many large-scale features, AIGFS demonstrates a significant reduction in tropical cyclone track errors at longer lead times.
- Efficiency:Efficiency being the most transformative feature AIGFS uses only 0.3% of the computing resources of the operational GFS for a single 16-day forecast. also finishes in approximately 40 minutes. This reduced latency means forecasters get critical data more quickly than they do from the traditional GFS.
- Area for future improvement: Though track forecasts are better, a degradation in tropical cyclone intensity forecasts is shown by v1.0, which future versions will address.
AIGEFS: This AI-based 31-member ensemble, similar to the GEFS Provides a range of possibilities for weather forecasters and decision-makers, rather than a single forecast model solution.
- Performance: Its forecast skill is comparable to the operational GEFS.
- Efficiency: requires only 9% of the computing resources of the operational GEFS.
- Area for future improvement: The ensemble’s ability to create a range of forecast outcomes is being improved continuously by the developers.
HGEFS: A 62-member “grand ensemble” created by combining the 31 members of the physical GEFS with the 31 members of the AI-based AIGEFS, the HGEFS is the most innovative application in the new suite. .
- Performance: A combination of two different modeling systems (one physics-based, one AI-based), the HGEFS creates a larger, more robust ensemble that more effectively represents forecast uncertainty. As a result, the HGEFS consistently outperforms both the GEFS and the AIGEFS across most major verification metrics.
- A NOAA first: NOAA, to our knowledge, is the first organization in the world to implement such a hybrid physical-AI ensemble system.
- Area for future improvement: HGEFS’s hurricane intensity forecasts.are being improved by the continuous work of NOAA.
A NOAA and industry-wide effort
A joint initiative between NOAA’s National Weather Service, Oceanic and Atmospheric Research labs, the Environmental Modeling Center in NOAA’s National Centers for Environmental Prediction, and the Earth Prediction Innovation Center has resulted in this initial model suite as an outgrowth of Project EAGLE.
“Using Project EAGLE and the Earth Prediction Innovation Center, NOAA scientists continue to work with members of academia and private industry on more advancements in forecasting technology,” added Jacobs.
Leveraging Google DeepMind’s GraphCast model as an initial foundation, The team further fine-tuned the model using NOAA’s own Global Data Assimilation System analyses. The Google model’s performance, particularly when using GFS-based initial conditions has improved by this additional training with NOAA data.
