Exploiting radar polarimetry for nowcasting thunderstorm hazards using deep learning

Abstract. Severe convective weather events, such as hail, lightning and heavy rainfall pose a great threat to humans and cause a considerable amount of economic damage. Nowcasting convective storms can provide warning signals and mitigate the impact of these storms. Dual-polarization weather radars are a crucial source of information for nowcasting severe convective events; nevertheless, they are most often not considered in nowcasting. These radars provide signatures of different hydrometeors. This work presents the importance of polarimetric variables as an additional data source for nowcasting thunderstorm hazards using an existing neural network architecture with convolutional and recurrent layers. This network has a common framework, which enables nowcasting of hail, lightning and heavy rainfall for lead times up to 60 min with a 5 min resolution. The study area is covered by the Swiss operational radar network, which consists of five operational polarimetric C-band radars. Results indicate that including polarimetric variables and quality indices improve the accuracy of nowcasting heavy precipitation and lightning, with the largest improvement found for heavy precipitation.