DiffCast: A Unified Framework via Residual Diffusion for Precipitation Nowcasting

TL;DR

Precipitation nowcasting is highly challenging due to chaotic evolution, requiring both accurate motion tracking and detailed local fluctuations. DiffCast tackles this by splitting forecasting into a global deterministic motion component and a local stochastic residual, modeled with a residual diffusion process conditioned on motion priors. The framework is backbone-agnostic, end-to-end trainable, and demonstrates state-of-the-art improvements across four radar datasets by leveraging a Global Temporal UNet and segment-level diffusion. This approach yields sharper, more accurate forecasts with realistic detail, offering practical benefits for meteorology and smart-city applications.

Abstract

Precipitation nowcasting is an important spatio-temporal prediction task to predict the radar echoes sequences based on current observations, which can serve both meteorological science and smart city applications. Due to the chaotic evolution nature of the precipitation systems, it is a very challenging problem. Previous studies address the problem either from the perspectives of deterministic modeling or probabilistic modeling. However, their predictions suffer from the blurry, high-value echoes fading away and position inaccurate issues. The root reason of these issues is that the chaotic evolutionary precipitation systems are not appropriately modeled. Inspired by the nature of the systems, we propose to decompose and model them from the perspective of global deterministic motion and local stochastic variations with residual mechanism. A unified and flexible framework that can equip any type of spatio-temporal models is proposed based on residual diffusion, which effectively tackles the shortcomings of previous methods. Extensive experimental results on four publicly available radar datasets demonstrate the effectiveness and superiority of the proposed framework, compared to state-of-the-art techniques. Our code is publicly available at https://github.com/DeminYu98/DiffCast.

Figures (13)

• Figure 1: We evaluate different models' performances in precipitation nowcasting. The deterministic model (b) can nicely capture the moving trends but tends to get blurry appearance and underestimate the high-value echoes. The probabilistic model (c) can capture the appearance details well, but the predicted rainfall positions are inaccurate. Our method (d) is able to generate accurate prediction with nice appearance details.
• Figure 2: The overview of our DiffCast framework for precipitation nowcasting is shown in (a). The DiffCast models the precipitation process from two perspective: deterministic component and stochastic component. The former accounts for predicting a global motion trend by a coarse forecast, while the latter aims to incorporate stochasticity with auxiliary-conditioned diffusion into the coarse forecast by residual. The sub-figure (b) indicates the computing flow of our framework for training and inference, respectively. The green, orange and blue rectangles represent, respectively, radar echos segment, output of deterministic predictor and residual segment for diffusion model.
• Figure 3: Illustration of our Global Temporal Unet (i.e., GTUnet).
• Figure 4: Performance changes against different lead time in terms of CSI, HSS and LPIPS. For a better vision, we only show the curves of SimVP, Earthformer and PhyDnet with or without our framework.
• Figure 5: A visual comparison example on a precipitation event from SEVIR. The results of Earthformer, MAE, ConvGRU are similar, which is put into Appendix due to the space limitation.