New AI weather model shows promise for more accurate forecasts

A new AI weather model called GenCast, developed by Google’s DeepMind, has shown significant improvements in forecasting accuracy compared to traditional models. It can predict weather up to 15 days in advance with better precision, especially in the first week. GenCast uses past weather data to simulate multiple possible future scenarios, enhancing the reliability of forecasts. It outperformed the European Centre for Medium-Range Weather Forecasts model in over 97% of tested variables, achieving faster results in under 10 minutes. Despite its advancements, GenCast faces challenges, particularly in predicting extreme weather events and tropical cyclone intensity. Researchers are optimistic that ongoing improvements and hybrid models combining AI with physics will address these issues in the future.