Diffusion-Based Refinement for Kilometer-Scale Probabilistic Precipitation Nowcasting
2026-08-31 • Machine Learning
Machine Learning
AI summaryⓘ
The authors developed a new method called exPreCast-ENS that improves rainfall forecasts by turning a 4 km resolution forecast into a more detailed 1 km probabilistic forecast. Their approach uses past radar data and existing forecasts to better correct errors and capture small-scale rain variations. Tested on the Korean Peninsula and in France, the method showed it could detect many missed heavy rain areas while keeping false alarms low. It also runs quickly, producing hourly forecasts in just a few seconds on a GPU.
nowcastingradar precipitationensemble forecastingdiffusion modelsprobabilistic forecastsspatial resolutionflash floodssystematic forecast errorGPU computing
Authors
Dohyun Park, Changhoon Song, Tengyuan Chang, Yoo-Geun Ham, Youngjoon Hong
Abstract
Localized extreme precipitation is a major trigger of urban flash floods and landslides, yet producing nowcasts that combine fine spatial detail with probabilistic uncertainty remains challenging. Here we introduce exPreCast-ENS, a conditional residual diffusion framework that transforms the deterministic 4 km radar nowcaster exPreCast into a 1 km probabilistic ensemble while correcting systematic forecast errors. Conditioning on both the forecast and preceding radar observations lets the ensemble-mean correct the baseline rather than perturb it, while members represent unresolved fine-scale variability. Over the Korean Peninsula, skill improves with ensemble size. In two high-impact events in 2023, a 30-member ensemble recovers 38-47% of heavy-rain pixels missed by exPreCast while retaining approximately 95% of its correct detections and alarming on under 1% of the pixels it correctly left clear. The method generates a 1-h forecast in 3.4 s on a single GPU and yields consistent improvements on the French regional MeteoNet radar dataset.