【作者】 李南；

【导师】 魏鸣；

【作者基本信息】 南京信息工程大学 ， 大气遥感科学与技术， 2011， 博士

【摘要】 本文针对多普勒天气雷达数据的质量控制和实际应用中需要解决的问题,进行了退速度模糊、风场反演和临近预报等方面的研究：1.针对目前业务应用中多普勒天气雷达资料退速度模糊容易退错的问题,提出了一种新的自动退模糊方法——零速线搜索法。从雷达原点开始到最远探测距离,搜索线性风场的两条零速线,同时确定零速线两侧区域的标准速度符号。然后,逐点比较径向速度场中数据点的速度符号与所属区域的速度符号,若两者相同,认为该数据点速度值正确；若两者相反,则认为该数据点为速度模糊,并修改为正确的速度值。实例的退模糊结果表明,零速线搜索方法合理有效,能够正确退除存在孤立回波、距离折叠等情况的速度模糊。从思路和方法上改进了传统线性外推方法退速度模糊容易退错的问题。2.针对多普勒雷达资料反演风场的VVP(Velocity Volume Processing)方法中出现的病态矩阵问题,提出了一种改进的VVP方法——SVVP(Step VVP)方法。SVVP方法采用分步反演的策略,计算线性方程组的各个变量,降低了方程组的条件数,克服了VVP方法中的病态矩阵问题,在分析体积很小的情况下仍可以得到较理想的反演结果,同时详细分析了反演过程中的误差来源和大小。对对流单体和台风资料的反演结果表明SVVP方法是可行的,能较有效地获得中小尺度对流系统的风场结构。3.针对利用多普勒天气雷达资料进行临近预报中需要改进对流单体的识别和跟踪的问题,提出了新的对流单体识别方法和预警方法,尝试将现代优化算法应用于对流单体追踪。新的对流单体识别方法利用搜索邻近点技术识别二维对流单体分量,并且改进了对流单体分量的垂直相关来构造三维对流单体,克服了传统方法中的识别缺陷。尝试利用现代优化算法(模拟退火算法,遗传算法,蚁群算法)进行对流单体的匹配和追踪,实验结果和理论分析显示,模拟退火算法和蚁群算法对于对流单体追踪简单有效,参数直观可调；遗传算法受限于遗传操作方式,效果不理想。基于风暴演化规律和特征量分布特征,将对流单体分为强对流单体和普通对流单体,并用支持向量机方法进行辨别,结果表明支持向量机能够指示对流单体的演化阶段和强弱变化,为强对流天气的预警提供了有力的工具和参考。更多还原

【Abstract】 Aiming at the quality control and problems that need to be solved in practical application, using the Doppler weather radar data in current operational application, this article makes studies on the velocity dealiasing, wind retrieval and nowcasting of Doppler weather radar data.1. To solve the problems that velocity dealiasing of Doppler weather radar data usually has mistakes in current operational application, a new automated velocity dealiasing method based on zero isodop searching has been developed. Zero isodops are searched point by point from the radar origin to the maximum detection range, and the accepted sign of each region separated by the zero isodops is determined. After that, the velocity sign at each gate is compared with the accepted sign of the region wherein the gate is. If they are consistent, the velocity is considered as true; otherwise as aliased and then dealiased to the true value. Dealiasing results on real cases indicate that the new algorithm is practicable and effective, especially on aliasing lack of references affected by discontinuous echo or range folding. It improves the velocity dealiasing problems, mistakes by using traditional linear extrapolation methods, in thoughts and methods.2. To solve the ill-conditioned matrix appeared in VVP (Velocity Volume Processing) wind retrieval, an improved VVP method, SVVP (step VVP) method, is proposed. SVVP method retrieves components of the wind field through a step wise procedure, which reduces the conditional number of equations and overcome the ill-conditioned matrix problems which currently limit the application of the VVP method. Variables can be retrieved even if the analysis volume is very small. In addition, the source and order of errors are analyzed in the retrieval. The improved method is applied to real cases including convective storms and typhoons, which show that it is robust and relative capability to obtain the wind field structure of the meso-scale convective system.3. To solve the storm identification and tracking problems in nowcasting by Doppler weather radar data, a new storm identification and warning technique is proposed, and modern optimization algorithms are used in storm tracking. The new identification method assembles contiguous storm points to constitute 2D storm components and improves the vertical association of storm components to construct 3D storms, which overcomes the deficiencies existing in traditional identification methods. Modern optimization algorithms (simulated annealing algorithm, genetic algorithm and ant colony algorithm) are tested to match and track storms. Experiment results and theoretical analysis show that simulated annealing algorithm and ant colony algorithm are effective and have intuitionally adjustable parameters, whereas results from genetic algorithm is unsatisfactory for the constraint of genetic operations mode. Based on the evolution properties and characteristic distributions, storms are specified as strong storms and general storms which then can be discriminated by a support vector machine (SVM). The performance of the SVM shows that it can indicate the development and evolution of a storm, providing an important aid in severe weather warning.更多还原