Microsoft has unveiled Aurora, a powerful AI model designed to dramatically improve environmental forecasting, including predictions of air quality, hurricanes, and typhoons. Built using over one million hours of atmospheric simulation data, Aurora can deliver global weather and pollution forecasts in under a minute. Its speed and scale allow it to generate ten-day predictions at a resolution of 0.1 degrees globally—comparable to the performance of leading traditional weather supercomputers, but significantly faster and more efficient.
In benchmark tests, Aurora has already demonstrated its capabilities by accurately forecasting events like Typhoon Doksuri’s landfall four days in advance. It also outperformed the U.S. National Hurricane Center in predicting the five-day paths of tropical cyclones during the 2022–2023 season. Microsoft emphasizes that Aurora is not just a technological achievement but a step toward using AI to enhance disaster preparedness, climate research, and global sustainability efforts through accessible and precise forecasting.
Leave a Reply