A lightning motion prediction technology based on spatial clustering method

Lightning warning is an essential means and a key technology for lightning disaster reduction. Passivity is the main drawback of existing lightning warning methods. This paper present a lightning prediction technology based on spatial clustering method to predict lightning motion by utilizing realtime and historical lightning monitoring data of lightning location system. On the contrary to the existing passive lightning prediction and warning methods, this technology initiatively predicts the prospective lighting area. The independent, discrete, random-happen flashes are aggregated to different thunderstorm groups similar to thunderclouds by the spatial clustering method presented in this paper. The position prediction of lightning can be achieved by tracking these thunderstorm groups and calculating the movement direction and speed of them. The lightning prediction method is applied to a certain region of central China. The experiment result shows that the predicted and real positions of lightning are approximately consistent. The average prediction accuracy is around 75%. Important support will be provided for improving the accuracy and reducing the false alarm rate of lightning warning by the presented technology.