Characteristics Analysis And Warning Of Lightning In XiangShan Region

With the development of the economy, great changes about people’s living environment have taken place. Now people pay more attention to the meteorological phenomenon which is closely connected with our life. In recent years, the climate in some places are abnormal, extreme weather events occur frequently, even one meteorological disaster would happen in two or more places at the same time. According to incomplete statistics, the average economic losses caused by meteorological disasters are more than10-50billion yuan every year in China, which has seriously affected the development of the society and people’s daily life. And thunder and lightning as underestimated meteorological disasters were paied more attention to. Xiangshan surrounded by two ports and two oceans is located in the southern of the Yangtze Delta. Xiangshan is made up of the east peninsula of Xiangshan and656islands and it is a typical island terrain. Its geographical location determines that Xiangshan has very obvious characteristics of subtropical monsoon climate and there are many thunderstorms in the summer. So understanding the characteristics of the activity of the thunderstorm and predicting the thunderstorm have significant influence on the lightning protection and disaster reduction.Atmospheric electric field instrument and lightning position indicator are the two main materials which can observe the activity of the thunder and lightning at present. When they are observing the activity, the two machines are affected by many factors such as the environment and so on. So when we use the two machines, we must analise the problem about them and then we can use the results from the observation to analyse the characteristics of the thunder and lightening in Xiangshan. On this basis, we put forward the prediction and warning about the thunder and lightning in Xiangshan.When the thunder strike the signal of the noise in the atmospheric electric field influences the analysis, collection and usage of the real atmospheric electric field. In this paper, the mean square error between the atmospheric electric field and the electric field without the signal of the noise is counted. Then we analyze how different wavelet function, thresholding method and wavelet decomposition level influence the denoising, and consider the signal to noise ratio during the lightning stroke. Considering lightning atmospheric electric field signal to noise ratio, it gets ideal de-noising effect when selecting sym5wavelet and db5wavelet combined with minimaxi and rigrsure threshold method, and wavelet decomposition level is6-7. This method is used to research the material about the Xiangshan’s atmospheric electric signal from2010to2012and realized the regular about the atmospheric electric signal during the lightning stroke. According to all these, we get the warning model parameters which will be used to predict the thunder and lightning.Using the approach refered to above to analyze and select the current intensity, the number of the lightning stroke during the return stroke, the intervals at the interstice of the return stroke, the distance between two adjacent sites, etc observed by the lightning position indicator from2010to2012in Xiangshan. It shows that the thunderstorms in every month correspond to the data recorded by man and the material handled by the atmospheric electric field instrument. And then we can draw the following conclusion. The average strength of negative C-G lightning in Xiangshan area is-24.975KA, which takes place of96.7%of the total area of flash. The average strength of positive CG lightning is38.775KA. The lightning mainly happens between March to October. It will obviously increase from June, in August, it comes to the peak, then if is about to decrease rapidly from September, which coincide the summer time. Lightning frequently mainly center on the time from11:00to20:00one day, and it will become the crest the value between14:00and17:00. The statistics about the steepness of lightning current waveform to the flash shows that the positive CG lightning’s steepness is above the negative CG lightning clearly after12:00. After8:00the steepness of the negative CG lighting chiefly concentrate on the figures between7and8KA/us. Besides, the lightning mostly happens in the port and the peaks of both areas in the whole Xiangshan area.Finally, we can use the data which is analyze and handle by the atmospheric electric field instrument and the lightning position indicator to establish the thunderstorm warning model based on the Bayes Discriminant in Xiangshan. We choose39samples(team A) to train the model and21samples (team B) to test the model. As a result, the POD of the model is80%, FRA of the model is27.3%and FOM of the model is20%. The phenomenon which happened in the mutation process of the atmospheric electric field appeared a half and one hour to two hours before the lightning stroke happened is added to the model. And we also used team A to train model and team B to test the model and the improved model’s POD and FOM remain unchanged, the FAR decreased to20%. And if team A and team B are all used to establish the model and train, the FAR of the improved model decreased obviously, but the POD also decreased. The FOM decreases the FAR by sacrificing the POD. In practical application, we should choose the predictor and establish the warning model according to the specific condition to reach the ideal effects.