Rain nowcasting with Radar reflectivity maps

Encadrants : Dominique Béréziat
Disponible : NON
Spécialité : IMA
Nombre d'étudiants : 3
Description :

Rainfall forecasting is an important environmental issue (flood anticipation, agriculture management…). Here we are interested in a short-time forecast (time horizon ranges from 30 minutes up to 2 hours). We use radar reflectivity maps which are robust to rainfall indicators. The literature on rain nowcasting using DL approaches is quite recent (2015). Up to 2021, the problem has been formulated as a classification task with less than 10 precipitation classes, such as no rain, light rain, moderate rain, etc.

Pré-requis : Pytorch.
Travail demandé : In this work, we want to address the rain nowcasting as a regression problem, and predict rain precipitation as a physical quantity and not as a class. This made the problem harder. To help, students are invited to draw inspiration from a dozen papers provided by the supervisor.
We use the Meteonet dataset and provide a ready to use dataloader.