Continuous Ensemble Weather Forecasting with Diffusion models

Weather forecasting has seen a shift in methods from numerical simulations to data-driven systems. While initial research in the area focused on deterministic forecasting, recent works have used diffusion models to produce skillful ensemble forecasts. These models are trained on a single forecasting step and rolled out autoregressively. However, they are computationally expensive and accumulate errors for high temporal resolution due to the many rollout steps. We address these limitations with Continuous Ensemble Forecasting, a novel and flexible method for sampling ensemble forecasts in diffusion models. The method can generate temporally consistent ensemble trajectories completely in parallel, with no autoregressive steps. Continuous Ensemble Forecasting can also be combined with autoregressive rollouts to yield forecasts at an arbitrary fine temporal resolution without sacrificing accuracy. We demonstrate that the method achieves competitive results for global weather forecasting with good probabilistic properties. Forecasting of physical systems over both space and time is a crucial problem with plenty of realworld applications, including in the earth sciences, transportation, and energy systems. A prime example of this is weather forecasting, which billions of people depend on daily to plan their activities. Weather forecasting is also crucial for making informed decisions in areas such as agriculture, renewable energy production, and safeguarding communities against extreme weather events.