基于云团的分组Z-R关系拟合方案效果评估与误差分析

Evaluation and Analysis of the Z-R Storm-Grouping Relationship Fitting Scheme based on Storm Identification

为了降低因Z-R关系不确定导致的雷达定量降水估测(Quantitative Precipitation Estimation,简称QPE)误差,提出了基于云团的分组Z-R关系拟合方案,在风暴单体识别算法得到的不同降水云团或同一个云团内部的不同数据分组区域内,拟合并采用不同的Z-R关系反演地面降水信息。以2013年6月5—7日的梅雨锋过程为例,使用覆盖长江中下游地区的28部多普勒雷达和全国逐分钟雨量计的观测资料,对单一动态关系、简单分组Z-R关系以及基于云团的分组ZR关系反演的雷达1 h QPE进行效果对比和误差分析,结果表明:(1)基于云团的分组Z-R关系可以有效识别降水云系的局部特征,这是基于云团的分组Z-R关系优于其他两种Z-R关系方案的重要原因。(2)雷达波束部分遮挡导致的偏弱反射率因子,对雷达QPE数据场的不连续性和Z-R关系的不确定性均有影响。(3)雷达硬件或雷达标定引入的偏强(弱)的反射率因子,与简单分组Z-R关系得到的雷达QPE局部高(低)估相关,这降低了简单分组Z-R关系在大范围降水过程中的适用性,但对基于云团的分组Z-R关系的影响较小。

A new grouping Z-R relationship fitting scheme is proposed based on the storm information derived by the storm-identification algorithm to reduce the error of quantitative precipitation estimation （QPE）, by which different Z-ll relationships are calculated and utilized respectively within separated storms identified by storm identification algorithm or data groups within the same storm to retrieve the surface rainfall information. Case study is presented by using the Meiyu front event between 05 and 07 June 2013, which covered the whole Mid- die-Lower Yangtze Plain. Using 28 S-band Doppler radar and 11623 minute-by-minute gauge observations, the unique dynamical Z-R re- lationship, simple grouping Z-R relationships and Z-R grouping relationships based on storm identifications are compared using radar I- hour QPE derived, respectively. The results show that （1） the storm-grouping relationship fitting scheme can effectively capture the local da- ta characteristics which make it superior to the other two fitting schemes. （2） The weaker radar echoes due to radar beam partial shielding not only accounts for the discontinuity of radar QPE, but affects the uncertainty of the Z-ll relationships. （3） The stronger or weaker radar e- choes result from by radar hardware or the calibration related to the overestimation or underestimation of radar QPE derived by the simple grouping Z-R relationships and make it unsuitable for the application in the large-scale rainfall event. But it has less influences on the storm-grouping relationships fitting scheme.