Scholarlink

Forecasting the occurrence of thunderstorms is a well-known challenge in weather prediction. Since a thunderstorm is by definition accompanied by at least one lightning, the goal is to forecast the occurrence of at least one lightning within a pre-defined area and a pre-defined time interval. This study reports on the development of a postprocessing method that is based on neural networks to translate the information from the lightning potential index (LPI) derived from the ICON-EU ensemble prediction system into a calibrated lightning probability. The ground truth from which the postprocessing method learns this translation function is given by lightning observations from the LINET network. The setup of the postprocessing method is presented together with results regarding the calibration and sharpness of the final product and the diurnal cycle of the predictions.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Forecasting Lightning Probabilities Derived from the Lightning Potential Index Using Neural Networks

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Manuel Baumgartner,
  • Guido Schröder,
  • Cristina Primo

摘要

Forecasting the occurrence of thunderstorms is a well-known challenge in weather prediction. Since a thunderstorm is by definition accompanied by at least one lightning, the goal is to forecast the occurrence of at least one lightning within a pre-defined area and a pre-defined time interval. This study reports on the development of a postprocessing method that is based on neural networks to translate the information from the lightning potential index (LPI) derived from the ICON-EU ensemble prediction system into a calibrated lightning probability. The ground truth from which the postprocessing method learns this translation function is given by lightning observations from the LINET network. The setup of the postprocessing method is presented together with results regarding the calibration and sharpness of the final product and the diurnal cycle of the predictions.