The Rise of AI: Outperforming Meteorologists in Weather Predictions

A machine learning-based weather prediction program called “GraphCast” predicts weather variables over the span of 10 days in under one minute. GraphCast outperformed traditional weather pattern prediction technologies at a 90% verification rate, according to scientists. The AI-powered weather prediction program takes in “the two most recent states of Earth’s weather,” including the variables from the time of the test and six hours prior. GraphCast can predict the state of the weather in six hours based on that data. The tool predicted the landfall of Hurricane Lee in Long Island 10 days before it happened, outperforming traditional weather prediction technologies. GraphCast can also predict severe weather events such as tropical cyclones and waves of extreme temperatures over regions. The algorithm can be re-trained with recent data, improving its ability to predict weather patterns and changes that align with climate change. Google may integrate GraphCast into its products, with a push for better storm modeling already paving a path for supercomputers in the space. The NOAA is working on developing models for more accurate readings on severe weather events and intensity forecasts for hurricanes.