Time-frequency Analysis Of The Atmospheric Electric Field Data And Application In Short-term Lightning Warning

The activities of the electric field is closely related with the solar activity, lightning activity, sandstorms and earthquakes, which represents an important characteristic parameters to the areas such as space physics, atmospheric physics and space environment, etc. Research on the electric field data in time and frequency domain under different weather conditions is helpful for us to analyze weather processes and forecast the first CG flash.The work is supported by China Commonweal Industry Research Project. Under the premise of using Jiangsu Province Lightning Detection information and Nanjing Doppler radar data as a supplement to ensure the time of the first CG flash, the paper mainly use the atmospheric electric field data to analyze the variation of the atmospheric electric field under different weather conditions. The conclusions are as follows:(1)The diurnal variation of fair-weather atmospheric electric field exhibit in two ways, one of which is a single oscillation with a peak and trough, and another is a double oscillation with two peaks and troughs. The result shows that the single waveform stations often locate in non-industrial towns or suburban. The first trough of all the stations have a good correspondence between02:00-06:00. The seasonal variation of these data is obviously higher in winter and lower in summer.(2)Thunderstorm development process usually contain three stages:towering cumulus stage, mature stage and dissipation stage. The variation of thunderstorm atmospheric electric field in all stations corresponds with the three stages; but there exists different time interval and electric intensity of different state of affairs.(3)Compared to time domain, the variation of atmospheric electric field has a more obvious feature in frequency domain, by examine the amount of the peak&trough of the characteristic harmonic and the amplitude range maybe regard as a new parameter in weather forecast of thunderstorms. After a large number of samples authentication, the data which choose30minutes before the first CG flash which calculated by FFT can be a standard spectrum to distinguish the two types of weather conditions. Meanwhile, using Euclidean distance method on unknown weather samples to classify the data and examine the results. With the extension of the forecast time, the TS score reduced from0.8to0.5, which shows that this new method can be regarded as a reference to the short-term and approaching first CG flash forecast.