Portability of Single-Station ML.Based Lightning Nowcasting Models

Significant damages and hazards caused by lightning can be avoided by predicting lightning occurrences and taking precautionary measures. In this study, a machine learning model was trained on the available data from twelve different weather stations in Switzerland. The model shows promising performance for predicting unseen data. Furthermore, the portability of the model is tested on the data recorded from a weather station in Suzhou, China, showing only a 5 percent drop in performance although the model has been trained on a different dataset and geographical location.