Characteristics Of Cloud-to-ground Flashes Of Two Thunderstorms Based On Accurate Strokes Counting Method

In this paper the continuous acquisition system of lightning electromagnetic field waveform based on the frequency of VLF/LF is used. By using anticipation program united with manual method to recognize the return stroke process accurately, we analyze the waveform records of two summer storms on Anhui region occurred at July9,2012and August19,2012. Multi-station TOA method and magnetic information is used to improve data accuracy. We analyze the characteristic parameters of negative and positive cloud-to-ground flashes in two storms and how they chang with the development of storm lifecycle. The method used in this paper is as accurate as the artificial method and also has the continuity of the multi-station network. This method is very useful for lightning observation. A summary of some interesting conclusions is summarized below:(1) To analyze the basic characteristics of negative cloud-to-ground flashes we find that there’re some common characteristics and multiplicity between the two different thunderstorms. Storm0709is a frontal thunderstorm and Storm0819is a local thunderstorm. There’re some common characteristics about these two thunderstorms. Such as the percentage of negative CG flashes (92.6%-86.6%), the mean value of inter-stroke intervals (75.3ms-61.8ms), the peak field ratio of subsequent strokes to first return strokes (0.49～0.52), the normalized peak field of return strokes (5.0V/m-5.2V/m), the process duration of preliminary breakdown pulses train and interval between preliminary breakdown pulses. These characteristics reflect the common features of negative CG flashes about the two thunderstorms. But there’re also some difference are worthy of our attention. The percentage of single stroke negative CG flashes of Storm0819(30.6%) is twice about which of Storm0709(15.5%). The mean number of strokes per flash of Storm0709(4.2) is bigger than Storm0819(3.3). The process duration of negative CG flashes of Storm0709(363.7ms) is longer than Storm0819(269.6ms). The mean number of strokes per flash is an important parameter to influence the process duration of negative CG flashes. The PB/RS value of Storm0709(0.25) is bigger than Storm0819(0.14) and the percentage of PB/RS0.1in two storms was14.9%and38.9%. This case may be caused by the difference between detection efficiency of the system for storms with different distance. So we need to consider the whole lifecycle of different thunderstorms and different thunderstorm days to get exacted characteristics of negative CG flashes.(2) By analyzing the EMF waveform records obtained by the system in a continuous manner, we found some characteristics of negative CG flashes variety with the development of the thunderstorm’s lifecycle. The lightning frequency of both thunderstorms has a’lightning jump’ in the mature stage and the positive CG flashes frequency of Storm0709increased intensely in the decaying stage. Mean number of strokes per flash variation with different thunderstorm stage of Storm0709is between3.0-5.3and of Storm0819is between1.9-4.1. Mean number of strokes per flash is sensitive to the ratio of single stroke flashes and they present negative correlation relationship. The variation trend of process duration of negative CG flashes is volatility and which is consisted with the change trend of mean number of strokes per flash in thunderstorm’s lifecycle. The mean value of inter-stroke interval of mature stage is smaller than other stages of the thunderstorm and which is sensitive to the intensity of the lightning discharge. So we need to consider different stages of thunderstorms to get accurate characteristics of CG flashes return strokes.(3) The mean value of inter-stroke intervals increases with return stroke order and decreases with return stroke number. The inter-stroke intervals probability distribution of Storm0709is similar to Storm0819and both of them obey the lognormal distribution. They are mainly concentrated in40-120ms and the maximum probability are60ms. The inter-stroke intervals and the percentage of intervals>100ms increase with the return stroke order. That means the percentage of long time intervals of return strokes increase with return stroke order. By analyzing the relationship of time intervals and return stroke order for negative CG flashes with different return stroke number we found the same characteristics. In fact that the charge neutralized by subsequent strokes are moved from the end of the lightning channel by K process, the occurrence of subsequent strokes is restricted to the charge environment of the thunder cloud. The mean value of inter-stroke interval for both storms decrease with return stroke number, which means the more strokes a flash has, the shorter time interval is. This phenomenon can be explained by the charge of thunder cloud and electrical conductivity of lightning channels.(4) We found that the peak field of return strokes decreased with return stroke order by analyzing the correlation with two different methods and the peak field of first return strokes increase with return stroke number per flash. The ratio of peak field of subsequent stroke to first return stroke (Es/Ef) is obey to logarithmic normal distribution for Storm0709and Storm0819. The value of Es/Ef decrease with return stroke order which means the peak field of return strokes is weaker and weaker with the return stroke order increasing. In general that the peak field of first return stroke is stronger than that of subsequent strokes, but the cases of peak field of subsequent strokes stronger than first return stroke is not rare. The peak field of single stroke flashes is weaker than which of first return stroke of multi-stroke flashes. By analyzing the relationship of peak field of strokes and inter-stroke intervals, we find that intervals less than50ms are dominated by the fact that the larger stroke occurs first, while intervals more than150ms are featured by the preceding return stroke being weaker than the subsequent one. The electrical conductivity of the lightning channel can influent the charge accumulation process and long intervals make the electrical conductivity of channel lessened. So there’re more time for thunder cloud to accumulate charge and trig strokes stronger than normal level.(5) Such characteristics as the process duration of preliminary breakdown pulses train (T<j), the time interval of pulses (Tj), ratio of peak field of biggest PB pulse to the first return stroke (PB/RS) and the time interval of PB and first return stroke (PB-RS) are analyzed in this paper. There’re1212negative CG flashes having PB belong to Storm0709and the value of Td and T; are2.3ms and194.2μs.275negative CG flashes with PB belong to Storm0819and the value of Td and Tj are2.7ms and176.8μs which are similar to Storm0709. PB/RS value of Storm0709is0.25and PB-RS value is23.2ms. The same characteristics of Storm0819is0.14and36.6ms. We summarized and compared the results obtained from different authors. Note that the means of the PB/RS ratio are relatively high in higher latitudes and as the latitude decreases the mean of the PB/RS ratio also decreases. Any differences in the PB pulse trains in different geographical regions may indicate differences in the initial breakdown processes in the clouds. It suggested that the reason for this difference is the weaker positive charge pocket in low latitudes thunderclouds compared to those observed in higher latitudes. We also find that the PB-RS separation is short at the high latitude region while having long durations in the low latitudes. This could be due to the higher altitude of charge centers in low latitude regions.(6) We summarized basic characteristics of positive CG flashes for two storms and further study needs more different positive CG flashes data to support. The percentage of positive CG flashes of Storm0709is7.5%and the ratio of single stroke flashes is93.3%. The maximum number of strokes per flash is2and the mean number of strokes per flash is1.1. The same characteristics of Storm0819are13.4%,84.5%,3and1.2which is similar to Storm0709. The percentage of positive CG flashes decreases with the total number of CG flashes increasing. The total flash number of Storm0709is three times about Storm0819and the percentage of positive CG flashes of Storm0709is lower than Storm0819. The mean time interval of strokes of Storm0709is92.1ms which is much higher than Storm0819(25ms). We find that the percentage of time interval>100ms of Storm0709(47.1%) is much higher than Storm0819(18.8%). This may be due to the different types of the two storms and the small simple size also has an impact on this phenomenon. We need to consider different types of thunderstorms and make large sample statistics to get the accurate characteristics of positive CG flashes.