三维雷达反射率资料用于层状云和对流云的识别研究

Identification of Stratiform and Convective Cloud Using 3D Radar Reflectivity Data

基于层状云和对流云的雷达反射率分布的三维形态特征,提出了识别层状云和对流云的6个候选识别参数,它们分别是:组合反射率及其水平梯度,反射率因子等于35 dBZ的回波顶高及其水平梯度、垂直累积液态水含量及其密度。通过分析候选识别参数分布图和选取的反射率垂直剖面图,用人机交互方式挑选“真实的”层状云和对流云区,统计这6个候选识别参数分布的概率密度特征;最后确定把分布概率密度更集中的组合反射率水平梯度、35 dBZ的回波顶高的水平梯度和垂直累积液态水含量密度作为识别参数,利用模糊逻辑法进行层状云和对流云的识别。用三个个例进行了识别试验,并把用模糊逻辑法识别的结果与用改进的巅峰值法识别的结果进行了比较,结果表明:用模糊逻辑法和改进的巅峰值法都能合理地识别大部分层状云和对流云;由于改进的巅峰值法只考虑了反射率分布的二维形态特征,它容易把对流核的外围识别成层状云,把厚实的层状云识别成对流云,而考虑了反射率分布的三维形态特征的模糊逻辑法在这两个方面有很大改善。

An automatic algorithm for the partitioning of radar reflectivity into convective and stratiforrn rain classifications has been developed and tested using volume scan radar reflectivity data from the Guangzhou and Fuzhou Doppler weather surveillance Radar. Based on the differences of radar reflectivity distribution morphology between convective and stratiforrn rain, six preparative reflectivity-morphological parameters are presented, which are composite reflectivity and its horizontal gradient, echo top height associated with 35 dBZ reflectivity and its horizontal gradient, vertically integrated liquid water content and its density. To arrive at a set of skillful separation parameters, the probability densities of the six separation attributes from ＂true＂ stratiforrn and convective rain are obtaineck Ideally, the statistics should come from a large sample in a site- and seasonality-specific manner, but large sample estimation of parameter statistics is not operationally viable. So the approach taken here is to estimate the parameter statistics from a small but very informative sample. Such a sample should contain, at least, a precipitation event with widespread, well-developed and clearly distinguishable areas of convective and stratiform precipitation. This paper uses the representative squall line on 22 March 2005 and mixed precipitation with bright band enhancement on 23 June 2005 at Guangzhou. Areas of ＇true＇ stratiform and convective precipitation are identified by analyzing images of six preparative parameters and selective vertical cross section of reflectivity in man-machine interaction manner. Results show that the densities of the horizontal gradient of composite reflectivity, the horizontal gradient of echo top height associated with 35 dBZ reflectivity and the density of vertically integrated liquid water content which are identified as the ultimate stratiform and convective cloud separation parameters are more concentrative than those of composite reflectivity, echo top height associated with 35 dBZ reflectivity and vertically integrated liquid water content, and the cross parts exist in the probability density function graphs of stratiform and convective cloud separation parameters. It is not very desirable to select a single set of thresholds for stratiform and convective cloud separation which could not possibly work well consistently and reliably for all sites, all seasons, and under varying conditions of radar calibration accuracy. Therefore, the fuzzy logic method is used for stratiform and convective cloud separation. The membership function of the fuzzy logic method is constructed according to the probability density features of the ultimate stratiform and convective cloud separation parameters, and the asymmetric trapezoidal membership function is chosen as the form of the membership functions. Three cases from Ghuangzhou and Fuzhou Doppler weather surveillance Radar are studied using the fuzzy logic method and the advanced maxima method. Results show that both the methods can separate most stratiform and convective rain. Because of using only the two-dimensional reflectivity morphology feature, the advanced maxima method has two main sources of misclassification： convective classification being assigned to heavy stratiform rain, and stratiform classification being assigned to the periphery of convective cores. By applying information based on the three-dimensional hydrometeor field inferred from radar reflectivity, the fuzzy logic method improves the performance of echo classification by correcting two main error sources of the advanced maxima method. Heavy stratiform rain and the periphery of convective cores are both classified correctly by the fuzzy logic method.