Downscaling weather forecasts from Low- to High-Resolution with Diffusion Models
This work provides a probabilistic downscaling method for weather forecasting, enabling high-resolution ensemble generation from coarse inputs, which is important for operational meteorology.
The authors introduce a diffusion-based method for global atmospheric downscaling that transforms low-resolution (100 km) ensemble forecasts into high-resolution (30 km) ensembles. The model improves probabilistic skill (FCRPS) for surface variables, reproduces target power spectra, captures multivariate relationships, and generates extreme values consistent with target ensembles in tropical cyclones.
We introduce a probabilistic diffusion-based method for global atmospheric downscaling implemented within the Anemoi framework. The approach transforms low-resolution ensemble forecasts into high-resolution ensembles by learning the conditional distribution of finer-scale residuals, defined as the difference between the high-resolution fields and the interpolated low-resolution inputs. The system is trained on reforecast pairs from ECMWF IFS, using coarse fields at 100 km to reconstruct fine-scale variability at 30 km resolution. The bulk of the training focuses on recovering small-scale structures, while fine-tuning in high-noise regimes enables the generation of extremes. Evaluation against the medium-range IFS ensemble target shows that the model increases probabilistic skill (FCRPS) for surface variables, reproduces target power spectra at small scales, captures physically consistent multivariate relationships such as wind-pressure coupling, and generates extreme values consistent with those of the target ensemble in tropical cyclones.