To improve the radar quantitative precipitation estimation,an optimization rainfall algorithm of S-band dual-polarization radar,named HCA-LIQ,based on hydrometeor identification is developed by referring to the Colorado State University (CSU)-ICE algorithm in this study.The radar estimator R (ZH),R (ZH,ZDR),R (KDP) calculated from the raindrop size distribution data collected in South China are used in this algorithm.Both the data collected from the S-band dual-polarization radar in Zhuhai,Guangdong Province and a network of rain gauges are used to evaluate the performance of the new algorithm.Comparison is also performed between the HCA-LIQ and CSU-ICE optimization algorithms and the traditional R (ZH) method.The results show that the HCA-LIQ optimization algorithm is well correlated with gauges and presents high stability.In addition,the distribution of hourly accumulation bias has light relation with the distance from the radar.The estimation results of the precipitation events show that two kinds of optimization algorithms are obviously superior to the traditional R (ZH) method for convective precipitation;the R (ZH) method is better than the two optimization algorithms for mixed cloud precipitation;the three errors statistics of the HCA-LIQ optimization algorithm are superior to the CSU-ICE algorithm.According to the bias statistics of the classification of rainfall intensity,the new HCA-LIQ optimization algorithm bias decreases by 23％ for light rain 71％ for heavy rain and 68％ for torrential rain respectively in comparison to the traditional R (ZH) method.%为了提高雷达定量降水估测的精度,本文利用雨滴谱数据、实际雨量以及不同偏振参量建立起的降水估测公式,参考CSU-ICE算法降水估测模型,建立了一种基于相态识别的S波段双线偏振雷达最优化定量降水估测算法(简称HCA-LIQ).利用广东珠海S波段双线偏振雷达数据和华南密集的地面自动站网的雨量,以两次华南夏季典型的降水过程为例,对该最优化算法的估测效果进行了统计分析,并进一步与CSU-ICE方法、传统的R(ZH)方法进行了比较.结果表明:以地面雨量计为标准,HCA-LIQ最优化算法表现出与雨量计较强的相关性且有着较好的稳定性.雷达相对地面雨量计小时雨量估测的偏差分布与离雷达的距离关系不大.按过程分类的结果显示,强对流云降水时,两种最优化算法要明显优于传统的R(ZH)方法;混合云降水时,最优化算法的效果并不比R(ZH)方法好;HCA-LIQ最优化算法比CSU-ICE算法效果更佳.按雨强分类统计时发现,与传统的R(ZH)方法相比,HCA-LIQ最优化算法对小雨的估测偏差降低了23％,对中雨的估测偏差相当,对大雨、暴雨的估测偏差分别降低了71％、68％.
展开▼
