Extreme Precipitation Nowcasting using Multi-Task Latent Diffusion Models

TL;DR

This paper introduces a multi-task latent diffusion model for extreme precipitation nowcasting, effectively capturing spatial details across varying intensities and outperforming existing methods on a benchmark dataset.

Contribution

The paper proposes a novel multi-task latent diffusion approach that decomposes radar images into intensity-based components for improved precipitation prediction accuracy.

Findings

Achieves 13-26% improvement in Critical Success Index (CSI).

Outperforms state-of-the-art precipitation prediction methods.

Effectively models multiple precipitation intensities separately.

Abstract

Deep learning models have achieved remarkable progress in precipitation prediction. However, they still face significant challenges in accurately capturing spatial details of radar images, particularly in regions of high precipitation intensity. This limitation results in reduced spatial localization accuracy when predicting radar echo images across varying precipitation intensities. To address this challenge, we propose an innovative precipitation prediction approach termed the Multi-Task Latent Diffusion Model (MTLDM). The core idea of MTLDM lies in the recognition that precipitation radar images represent a combination of multiple components, each corresponding to different precipitation intensities. Thus, we adopt a divide-and-conquer strategy, decomposing radar images into several sub-images based on their precipitation intensities and individually modeling these components. During the prediction stage, MTLDM integrates these sub-image representations by utilizing a trained latent-space rainfall diffusion model, followed by decoding through a multi-task decoder to produce the final precipitation prediction. Experimental evaluations conducted on the MRMS dataset demonstrate that the proposed MTLDM method surpasses state-of-the-art techniques, achieving a Critical Success Index (CSI) improvement of 13-26%.