A study of thunderstorm characteristics in Northwest Russia using neural networks
a:2:{s:4:"TEXT";s:76:"A.A. Sin’kevich, A.B. Kurov, Y.P. Mikhailovskii, M.L. Toropova, N.E. Veremei";s:4:"TYPE";s:4:"text";}
The Voeikov Main Geophysical Observatory, St Petersburg, Russia
Keywords: thunderstorm, lightning, radar characteristics, regression analysis, neural networks
Abstract
The paper presents the results of analysis of radar characteristics of clouds, including the polarization ones, and lightning data for June 9 2020, in the vicinity of Saint-Petersburg. The characteristics of thunderstorm and clouds without lightning are compared. Statistical difference between two groups of clouds has been found. The regression analysis of the correlation between the lightning flash rate and radar characteristics of clouds is performed using neural networks. The impact of these parameters on lightning flash rate has been estimated. A mathematical equation for calculating the lightning flash rate using the differential reflectivity maximum and the cloud volume with the reflectivity equal to 35 dBZ is derived.
|
