{"ID":2899010,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2507.01354","arxiv_id":"2507.01354","title":"Efficient Kilometer-Scale Precipitation Downscaling with Conditional Wavelet Diffusion","abstract":"Effective hydrological modeling and extreme weather analysis demand precipitation data at a kilometer-scale resolution, which is significantly finer than the 10 km scale offered by standard global products like IMERG. To address this, we propose the Wavelet Diffusion Model (WDM), a generative framework that achieves 10x spatial super-resolution (downscaling to 1 km) and delivers a 9x inference speedup over pixel-based diffusion models. WDM is a conditional diffusion model that learns the learns the complex structure of precipitation from MRMS radar data directly in the wavelet domain. By focusing on high-frequency wavelet coefficients, it generates exceptionally realistic and detailed 1-km precipitation fields. This wavelet-based approach produces visually superior results with fewer artifacts than pixel-space models, and delivers a significant gains in sampling efficiency. Our results demonstrate that WDM provides a robust solution to the dual challenges of accuracy and speed in geoscience super-resolution, paving the way for more reliable hydrological forecasts.","short_abstract":"Effective hydrological modeling and extreme weather analysis demand precipitation data at a kilometer-scale resolution, which is significantly finer than the 10 km scale offered by standard global products like IMERG. To address this, we propose the Wavelet Diffusion Model (WDM), a generative framework that achieves 10...","url_abs":"https://arxiv.org/abs/2507.01354","url_pdf":"https://arxiv.org/pdf/2507.01354v1","authors":"[\"Chugang Yi\",\"Minghan Yu\",\"Weikang Qian\",\"Yixin Wen\",\"Haizhao Yang\"]","published":"2025-07-02T04:46:28Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"physics.ao-ph\"]","methods":"[\"Diffusion Model\"]","has_code":false}