Generative AI to quantify uncertainty in weather forecasting

A generative AI model called SEEDS has been developed by Google to efficiently generate ensembles of weather forecasts, quantifying the uncertainty in weather conditions. This technology provides a cost-effective solution compared to traditional physics-based forecasting models. SEEDS leverages generative AI to produce real-weather-like forecasts that match or exceed physics-based ensembles. It allows for the generation of large ensembles to accurately assess rare but high-impact weather events.