Articles | Volume 13, issue 6
Model description paper
 | Highlight paper
11 Jun 2020
Model description paper | Highlight paper |  | 11 Jun 2020

RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting

Georgy Ayzel, Tobias Scheffer, and Maik Heistermann

Data sets

RYDL: the sample data of the RY product for deep learning applications G. Ayzel

RainNet: pretrained model and weights G. Ayzel

Model code and software

hydrogo/rainnet: RainNet v1.0-gmdd G. Ayzel

Short summary
In this study, we present RainNet, a deep convolutional neural network for radar-based precipitation nowcasting, which was trained to predict continuous precipitation intensities at a lead time of 5 min. RainNet significantly outperformed the benchmark models at all lead times up to 60 min. Yet, an undesirable property of RainNet predictions is the level of spatial smoothing. Obviously, RainNet learned an optimal level of smoothing to produce a nowcast at 5 min lead time.