模糊逻辑法在双线偏振雷达识别降水粒子相态中的研究

A Study of Fuzzy Logic Method in Classification of Hydrometeors Based on Polarimetric Radar Measurement

根据不同相态降水粒子的散射和空间取向等特征建立了基本形式为不对称的T型函数的模糊逻辑识别的隶属函数,完善了该识别方法的流程,并利用美国KOUN雷达的观测资料,详细讨论了利用双线偏振雷达观测资料识别降水粒子相态的方法,并对其在实际业务运用中的合理性和可行性进行分析探讨。通过分析,作者认为:(1)利用模糊逻辑方法处理双线偏振雷达测量到的偏振参数,可以识别降水区域的降水粒子相态,反映降水区域的相态结构,识别的结果基本合理,但还需要资料作进一步的研究。(2)从实际资料的分析结果来看,虽然利用模糊逻辑法识别降水粒子相态得到的结果比较粗,没有明确的数据特征值,但是它基本上能反映各种降水粒子的相态结构,其识别的结果对目前日常的天气预报参考和人工影响天气的作业指挥和效果评估来说还是符合可用的。为中国国内未来的双线偏振雷达业务运行提供参考和帮助。(3)根据有限的观测资料分析表明,降水区域中出现60dBZ以上的回波强度,并不一定就会出现冰雹,还有可能是大粒子的液态降水。综合考虑双线偏振雷达的测量参数,可以得到更合理的结果。

The polarimetric radar measurements are sensitive to the types, shapes, and sizes distribution. It is a useful remote sensing tool for the study of hydrometeor classification. A fuzzy logic system for classification of hydrometeor type based on polarimetric radar measurements is described. First, the membership function of fuzzy logic method is constructed according to the feasibility of scattering and orientation of different kinds of hydrometeors, and the asymmetric trapezoidal membership function is chosen as the form of the membership functions. Then, the method of classification of hydrometeors based on polarimetric radar measurement is discussed in detail by using KOUN radar observation data. The feasibility and rationality of its application in operation are also analyzed. The main conclusions are got as follows： （1） Fuzzy logic method based on polarimetric radar data can be used in hydrometeors classification, and the results are basically reasonable based on two case studies. It can reveal the development course of weather system, and classify hydrometeor types, but still need be further studied. （2） Although the classification result is coarse and has no definite value, it still can be used to initialize meso-scale numerical weather prediction models, offer reference to daily weather forecast and evaluate the results of artificial weather modification. It can provide reference in the polarimetric radar operation of China. （3） It may be not hail but big rain drops although the value of reflectivitv is larger than 60 dBZ based on data analysis.