Allen Institute releases HiRO-ACEv2 for high-resolution weather simulation
HiRO-ACEv2 generates 3 km resolution global climate data at approximately 1/300th the cost of traditional high-resolution atmospheric models.
Summary
The Allen Institute of AI has introduced HiRO-ACEv2, an updated climate emulation system designed to simulate extreme weather events. The system uses a two-part architecture: ACE, which operates at a 100 km resolution, and the new HiROv2 model, which downscales those outputs to 3 km detail. This approach allows for the reproduction of intense rain, wind, and pressure patterns with significantly lower computational requirements than models like X-SHiELD.
Technical improvements in HiRO-ACEv2
- HiROv2 downscaling model: Predicts multiple high-resolution fields simultaneously (rain, wind, and pressure) rather than just rainfall.
- Accuracy: Matches high-resolution statistics of X-SHiELD within 3%, even for extremely rare events.
- Extreme weather capability: Accurately estimates tropical cyclone energy and generates realistic hurricane structures, including eyes and rain bands.
- Efficiency: Reduces global data production costs to roughly 1/300th of X-SHiELD, with even higher savings for regional simulations.
- Architecture: Combines a coarse 100 km ACE model with the fine-scale HiROv2 model.
Why it matters
The release provides researchers with an inexpensive method to generate long-term, high-resolution climate data for studying extreme weather patterns.
Original reference: huggingface.co →