CasCast: Skillful High-resolution Precipitation Nowcasting via Cascaded Modelling
This work addresses precipitation nowcasting for disaster management, offering incremental improvements in extreme event prediction.
The paper tackles the challenges of modeling complex precipitation system evolutions and accurately forecasting extreme precipitation in high-resolution nowcasting, proposing CasCast, a cascaded framework that achieves competitive performance and surpasses baselines by up to +91.8% for regional extreme-precipitation nowcasting.
Precipitation nowcasting based on radar data plays a crucial role in extreme weather prediction and has broad implications for disaster management. Despite progresses have been made based on deep learning, two key challenges of precipitation nowcasting are not well-solved: (i) the modeling of complex precipitation system evolutions with different scales, and (ii) accurate forecasts for extreme precipitation. In this work, we propose CasCast, a cascaded framework composed of a deterministic and a probabilistic part to decouple the predictions for mesoscale precipitation distributions and small-scale patterns. Then, we explore training the cascaded framework at the high resolution and conducting the probabilistic modeling in a low dimensional latent space with a frame-wise-guided diffusion transformer for enhancing the optimization of extreme events while reducing computational costs. Extensive experiments on three benchmark radar precipitation datasets show that CasCast achieves competitive performance. Especially, CasCast significantly surpasses the baseline (up to +91.8%) for regional extreme-precipitation nowcasting.