【作者】 曾庆锋；

【导师】 王振会；

【作者基本信息】 南京信息工程大学 ， 大气遥感科学与技术， 2013， 博士

【摘要】 对大气电场观测实验区(南京、镇江、扬州以及常州地区)大气电场仪联网观测数据进行质量控制和统一标定,结合天气雷达、闪电定位仪和探空等多源资料对地面大气电场时空演变特征和电荷结构进行分析,提出基于单一来源资料及综合利用多源资料进行雷电临近预警的多种方法,对预警效果进行评估。论文主要工作和结论如下：(1)针对实验区大气电场联网观测数据,提出了地面大气电场仪单站观测数据的质量控制方法和联网观测数据的统一标定方法,实现了对联网观测数据的质量控制和客观分析。对统一标定之后的大气电场仪联网观测数据进行客观分析,显现出电场强度分布、雷达回波和闪电发生位置之间有较好的吻合性,发现闪电通常发生在电场梯度较大之处。(2)对不同天气特别是雷暴天气情况下的地面大气电场变化特征进行总结和分析,发现单体雷暴地面电场演变过程一般比较清晰而多单体雷暴地面电场演变过程则更为复杂；雷暴地面电场可分为正向增大、负向增大和上下震荡三种类型；大气电场演变过程与闪电的发生具有密切关系,闪电的发生演变过程和雷暴电场的三种类型具有较好的相关性。大气电场幅值慢变化、电场跳动和电场极性反转的现象是雷暴发生时大气电场的显著变化特征,这些特征可以用在雷电预警工作中。(3)建立了基于大气电场、雷达以及探空等数据的雷暴云电荷结构分析模型,研究了南京周边地区雷暴云电荷结构特征。假设雷暴云电荷团内电荷密度在空中呈准正态分布,对不同电荷结构(单极性、偶极性、三极性)及电荷团内电荷分布参数的雷暴云地面电场进行模拟计算,结合雷达回波的强度与空间分布数据、实测地面大气电场和模拟电场对比来反推雷暴云电荷结构。发现南京及周边地区雷暴电荷结构具有多样性和复杂性,但主要为偶极性,也存在三极性电荷结构。其次将雷达回波和雷暴初生特征以及环境温度和电荷区域特征运用于确定雷暴云电荷团内电荷密度为准正态分布的电荷分布参数(中心密度、中心位置、分布范围),得到了这些雷暴云过程的电荷密度随时间的演变过程以及电荷密度空间分布特征。同时将模拟得到的联网电场和电荷密度分布与插值联网电场以及其他观测资料进行对比分析发现,雷达回波和闪电发生位置与模型模拟结果基本吻合,而与插值联网电场有一定偏差。这一定程度说明模型以及结果具有一定合理性,也同时说明了站点布置对于电场联网的重要性。(4)提出了基于单一来源资料及综合利用雷达、大气电场仪、闪电定位系统及探空等多源资料进行雷电预警的方法,并对预警效果进行评估。单独利用大气电场、雷达和闪电定位等数据进行雷电临近预警并进行效果分析发现：利用闪电位置与电场测站之间的距离进行预警的方法是可行的；在3个雷达体扫时间内出现“电场绝对值超过1kv/m"的时间比例达1/3、在2个雷达体扫时间内出现“电场差分幅值绝对值超过0.15kV/m"达2次,这两个指标可分别用于南京以及周边地区的电场幅值阈值和电场差分阈值预警。将大气电场结合雷达数据对雷暴云首次地闪进行预报研究,发现6km高度上雷达反射率因子达到30dBZ是最优雷达预警阈值,两年数据统计得到POD为82.7%、FAR为14.6%、CSI为72.4%,平均预警时间是20.8min。同时,根据电场仪、雷达以及闪电定位仪观测的时间和空间关系,提出了将三种数据融合进行雷电预警的方法,得到POD为87.1%、FAR为16.5%、CSI为74.4%,平均预警时间为24.4min。对移过测区的典型雷暴过程进行区域性雷电临近预警,直观呈现局部测区的雷暴云和闪电的发生发展以及预报过程。同时区域性雷电临近预警可以增强大气电场仪的功能,不但能实现对一个地区进行预警,而且能对经过观测区的雷暴云的整个过程进行监测预警。更多还原

【Abstract】 In this paper, firstly, the data of atmospheric electric field in Nanjing, Zhenjiang, Yangzhou and Changzhou have been quality controlled, uniformed calibrated and network organized, then the spatial variations of atmospheric electric field and charge structures have been analyzed by combining with the data of radar, lightning location and sounding data. Finally, the lightning forecasting methods are proposed and the results have been analyzed by using the above multivariate observations separately and synthetically.According to the characteristcs of atmospheric electric field observations, quality controlling method for the data of atmosphere electric field has been proposed in order to improve the quality of atmospheric electric field data. After quality controlling, uniformed calibration and network organization, we ananlyed the data and found that the distribution characteristics of atmospheric electric field consistent with radar echo and lightning positionning data. The lightnings occur in the area where the electric field gradient is large.The characteristics of electric field under different weather conditions (especially thunderstorm) were analyzed. The results showed that the isolated thunderstorms’electric field had a clear evolution process while the multi-units thunderstorms’were complicated. The evolutions of thunderstorms’s electric field can be divided into three types:positive increasing, negative increasing and up-down jitter. In the meantime, the evolution of electric field closely realted with lighting occurrence and the lighting process was agreed well with three types of the evolution of thunderstorms’s electric field. The slow change, jumpiness and polarity inversion are the features of electric field used for distinguishing thunderstorm and can be used for lightning forecasting.As for the charge structure of thunderstorm, first, the electric field evolution was simulated under different charge structures, and then the charge structure of thunderstorm was analyzed by comparing the electric field observation data with simulation electric field with the help of radar data. It is found that the charge structures of thunderstorms are complicated and diversified. The charge structure of thunderstorms is mainly dipole and the tripole charge structure is also existed. The relation of radar reflectivity and thunderstorm initial feature, the characteristics of environmental temperature and charge distribution were implemented in the inversion of charge distribution parameters in this paper, and then a quasi-normal charge distribution model was proposed. The method was used to analyze three isolated thunderstorms and a MCS that passed through the experiment site in2009. The evolution of charge density with time and the space distribution of these thunderstorms can be abtained. The simulated space electric field and charge density space distribution were compared with other observation and networked electric field with interpolation method. It is found that the distribution of atmospheric electric field agrees well with radar echo and lightning location data. The lightning occurred in the place where the electric field gradient is large. In addition, it illustrate that the model and obtained parameters are reasonable to some extent, and the method for networking electrostatic field using radar and sounding data is feasible. There existed some deviation between networked electric field with interpolation method and the location of lightning and radar echo. Therefore, it also showed that the site deploy of electric field mill is very important.The methods and the effect of lightning forecasting based on atmospheric electrostatic field, radar and lightning location data separately were analyzed and found that:the lightning forecasting made use of lightning location data is feasible; the proportion of EF absolute amplitude which exceed1kV/m reach1/3is considered the best electrostatic field amplitude index, while the best electrostatic field differential index is defined as the times of difference absolute amplitude which exceed0.15kV/m reach2times in two radar scan times, which are more suitable indexes for the region in this study.Based on the lightning forecasting with atmospheric electrostatic field, radar and lightning location data separately, a method of cloud-ground (CG) lightning nowcasting based on electric field observations, radar data was proposed. The experimental (2009) and independent test (2010) results shows that POD (Probability of Detection, POD) is82.7%, FAR (False Alarm Rate, FAR) is14.6%, CSI (Critical Success Index, CSI) is72.4%and average lead time is20.8min. Besiades, on the basis of the temporal relationship of electric field mills, radar and lightning location mill observation when thunderstorms move toward, a lightning nowcasting method based on a combination of lightning location system, electric field and radar data was proposed and the results show that POD, FAR and CSI values were87.1%,16.5%and74.4%respectively and average lead time was24.4min. Finally, a classical thunderstorm example of regional lightning warning in the region is analyzed. Results show that this method can present a visual development of thunderclouds and lightning; it can provide some useful information about the region’s lightning characteristics.更多还原