Diagnosing upward lightning from tall objects from meteorological thunderstorm environments

Tall structures cause upward lightning (UL). It endangers and potentially damages tall wind turbines, which are a mainstay in the drive to reduce greenhouse gas emissions. Ground-truth lightning current measurements at the instrumented Gaisberg Tower (Austria) show that UL has no distinct annual cycle and that a conventional lightning location system (LLS) misses about half of UL events. The paper explores the possibility to diagnose UL only from larger-scale meteorological conditions of the thunderstorm environment in order to augment LLS data for damage risk assessment. It finds that wind-field thunderstorm environments account for nearly all UL in transition and cold seasons and still half in summer thus explaining the uniform annual UL distribution. A random forest model successfully diagnoses UL at the Gaisberg Tower using only meteorological variables. It performs best in winter and transitional seasons. Case studies show that the Gaisberg-trained model also succeeds in diagnosing UL at wind turbines in other regions, thus pathing the way for a better risk assessment of UL at wind turbines.