RainPro-8: An Efficient Deep Learning Model to Estimate Rainfall Probabilities Over 8 Hours
Rafael Pablos Sarabia, Joachim Nyborg, Morten Birk, Jeppe Liborius Sjørup, Anders Lillevang Vesterholt, Ira Assent
TL;DR
RainPro-8 addresses the challenge of high-resolution probabilistic precipitation forecasting over an 8-hour horizon in Europe by fusing radar, satellite, and NWP data within a compact architecture. It introduces an ordinality-aware training paradigm, an ordinal-consistent loss, and single-pass predictions to generate all lead times simultaneously, outputting probability maps for precipitation intensities across time and space. The approach yields substantial improvements over operational NWP, extrapolation-based methods, and prior deep-learning nowcasting models, while delivering improved inference speed and coherent uncertainty representations. The work further validates the method via probability maps, attribution analyses, and a radar-only SEVIR benchmark variant, highlighting practical benefits for weather risk management and decision support.
Abstract
We present a deep learning model for high-resolution probabilistic precipitation forecasting over an 8-hour horizon in Europe, overcoming the limitations of radar-only deep learning models with short forecast lead times. Our model efficiently integrates multiple data sources - including radar, satellite, and physics-based numerical weather prediction (NWP) - while capturing long-range interactions, resulting in accurate forecasts with robust uncertainty quantification through consistent probabilistic maps. Featuring a compact architecture, it enables more efficient training and faster inference than existing models. Extensive experiments demonstrate that our model surpasses current operational NWP systems, extrapolation-based methods, and deep-learning nowcasting models, setting a new standard for high-resolution precipitation forecasting in Europe, ensuring a balance between accuracy, interpretability, and computational efficiency.