{"ID":3005565,"CreatedAt":"2026-06-03T03:09:48.883664427Z","UpdatedAt":"2026-06-04T06:21:04.369492701Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2606.02661","arxiv_id":"2606.02661","title":"Learning to Refine: Spectral-Decoupled Iterative Refinement Framework for Precipitation Nowcasting","abstract":"Accurate precipitation nowcasting is vital for disaster mitigation, but deep learning methods face a key trade-off: regression models produce over-smoothed, spectrally decaying predictions that blur convective details and violate turbulence power laws; diffusion models generate realistic yet unanchored hallucinations lacking physical grounding. We propose Spectral-Decoupled Iterative Refinement (SDIR), a deterministic framework that reformulates nowcasting as progressive frequency-decoupled refinement. SDIR first extracts a stable low-frequency synoptic skeleton, then iteratively refines high-frequency textures under physical constraints, eliminating both blurring and hallucinations. It features a dual-path design: the Synoptic Frequency-Guided Former (SFG-Former) with Scale-Adaptive Transformers for global structure, and the Fourier Residual Refiner (FR-Refiner) with Scale-Conditioned Fourier Neural Operators for fine residuals. A Physically Consistent Power Spectral Density (PCPSD) loss with dynamic masking enforces a turbulence-consistent spectral distribution. Experiments on three benchmarks show SDIR significantly outperforms SOTA methods in spatial accuracy while achieving spectral fidelity competitive with diffusion-based methods, enabling reliable high-resolution operational nowcasting. Code link: https://github.com/RuntimeWarning/SDIR.","short_abstract":"Accurate precipitation nowcasting is vital for disaster mitigation, but deep learning methods face a key trade-off: regression models produce over-smoothed, spectrally decaying predictions that blur convective details and violate turbulence power laws; diffusion models generate realistic yet unanchored hallucinations l...","url_abs":"https://arxiv.org/abs/2606.02661","url_pdf":"https://arxiv.org/pdf/2606.02661v1","authors":"[\"Yunlong Zhou\",\"Chen Zhao\",\"Danyang Peng\",\"Fanfan Ji\",\"Xiao-Tong Yuan\"]","published":"2026-06-01T07:34:30Z","proceeding":"eess.IV","tasks":"[\"eess.IV\",\"cs.AI\",\"cs.LG\"]","methods":"[\"Diffusion Model\",\"Transformer\"]","has_code":false,"code_links":[{"ID":612741,"CreatedAt":"2026-06-03T03:09:48.883664427Z","UpdatedAt":"2026-06-03T03:09:48.883664427Z","DeletedAt":null,"paper_id":3005565,"paper_url":"https://arxiv.org/abs/2606.02661","paper_title":"Learning to Refine: Spectral-Decoupled Iterative Refinement Framework for Precipitation Nowcasting","repo_url":"https://github.com/RuntimeWarning/SDIR","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}