Universal Diffusion-Based Probabilistic Downscaling

arXiv:2602.11893v3 Announce Type: replace Abstract: We introduce a universal diffusion-based downscaling framework that lifts deterministic low-resolution weather forecasts into probabilistic high-resolution predictions without any model-specific fine-tuning. A single conditional diffusion model is trained on paired coarse-resolution inputs (~25 km resolution) and high-resolution regional reanalysis targets (~5 km resolution), and is applied in a fully zero-shot manner to deterministic forecasts from heterogeneous upstream weather models. Focusing on near-surface variables, we evaluate probabilistic forecasts against independent in situ station observations over lead times up to 90 h. Across a diverse set of AI-based and numerical weather prediction (NWP) systems, the ensemble mean of the downscaled forecasts consistently improves upon each model's own raw deterministic forecast, and substantially larger gains are observed in probabilistic skill as measured by CRPS. These results demonstrate that diffusion-based downscaling provides a scalable, model-agnostic probabilistic interface for enhancing spatial resolution and uncertainty representation in operational weather forecasting pipelines.

Related

• Achieving Skilled and Reliable Daily Probabilistic Forecasts of Wind Power at Subseasonal-to-Seasonal Timescales over France
• TopoFlow: Topography-aware Pollutant Flow Learning for High-Resolution Air Quality Prediction
• Barrier-enforced multi-objective optimization for direct point and sharp interval forecasting
• Generative 3D Gaussian Splatting for Arbitrary-ResolutionAtmospheric Downscaling and Forecasting

Source: arXiv cs.LG | 2026-04-21