HiRO-ACEv2: Simulating extreme weather at 3 km resolution for a fraction of the cost

Introducing HiRO-ACEv2, an advanced system that can generate high-resolution 3 km climate data at a fraction of the cost of other models. This update to the High Resolution Output Ai2 Climate Emulator (HiRO-ACE) system improves the accuracy of predicting rare weather phenomena like extreme rain, wind, and pressure events. HiRO-ACEv2 achieves this by combining ACE, an AI model that operates at a coarser 100 km resolution, with HiROv2, a downscaling model that can now predict multiple high-resolution fields simultaneously.

The new HiROv2 model has been trained to reproduce intense weather events with high accuracy, matching high-resolution statistics to within 3% even for rare events. It also estimates tropical cyclone energy accurately and generates hurricanes with realistic structures. When used in conjunction with an updated version of ACE, HiRO-ACEv2 can generate a full year of global 3 km data, closely matching the statistics of the high-resolution X-SHiELD model.

This system offers a cost-effective way for researchers to produce high-resolution data for extreme weather events, with potential savings of over a hundred times when focusing on a specific region. The team is also working on a faster version of HiROv2 and plans to expand its capabilities to handle different climates and additional weather variables like temperature and snow.