2021年郑州“7·20”极端暴雨雨滴谱特征及其对雷达定量降水估测的影响

利用雨滴谱仪观测的雨滴谱数据，分析了2021年7月20日郑州极端暴雨的雨滴谱特征，并结合双偏振雷达观测，分析了不同定量降水估测（QPE）方法在此次极端暴雨过程中的性能。结果表明，在此次极端暴雨过程的最强降水时段，雨滴谱表现为很高的粒子数浓度和很大的粒子平均直径；而整个降水过程雨滴谱的截距参数与我国其它地区雨滴谱特征差异不明显，但质量加权平均直径大于其他地区的雨滴谱；在降水最强的2021年7月20日08:00～09:00（协调世界时，下同）前后，雨滴谱的特征发生了显著变化，首先是质量加权平均直径迅速增长，随后粒子数浓度也陡增，从而导致降水率的迅速增强。使用郑州双偏振雷达数据，基于各种QPE方法和参数计算得到了08:00～09:00的雷达反演降水量，并与雨量计观测结果比较。结果表明对于基于反射率的QPE关系（R(Z_H)），如果不提高或者去除反射率上限进行QPE，会导致降水严重低估，且该方法对参数的准确性较为依赖；基于差传播相移率的QPE关系（R(K_dp)）对雨滴谱差异性敏感度相对较低，其性能主要依赖于差传播相移率的准确性；最优的R(K_dp)关系反演效果比R(Z_H)更好，能达到实际降水量的70%以上。

Raindrop Size Distribution Characteristics of the Extreme Rainstorm Event in Zhengzhou 20 July, 2021 and Its Impacts on Radar Quantitative Precipitation Estimation

In this study, we investigated the raindrop size distribution (DSD) characteristics of an extreme rainstorm event on July 20, 2021, in Zhengzhou using disdrometer observation data. The performance of many radar quantitative precipitation estimation (QPE) approaches was then examined using polarimetric radar data. The results show that during the peak rain rate period, the DSD indicated a high number concentration and large mean particle size. During this event, the normalized intercept parameter was similar to those observed in other regions of China, but the mass-weighted diameter was significantly higher. The DSD experienced significant changes before the peak rain rate period. The number concentration increased as the mass-weighted diameter increased, resulting in a quick increase in the rain rate. Polarimetric radar data were used to calculate the hourly QPE rainfall during 0800 UTC–0900 UTC, July 20, 2021, based on several QPE approaches and parameters, and then the performances of each approach were examined against the gauge observation. The result showed that the cap of reflectivity-based estimator (R(Z_H) )should be removed or raised; otherwise, the rainfall would be significantly underestimated. The estimator was sensitive to the QPE parameters; however, the specific differential phase-based estimator (R(K_dp)) was relatively insensitive to the QPE parameters, and the accuracy of the specific differential phase was responsible for its performance. The best R(K_dp) estimator reached over 70% of the observational hourly rainfall and outperformed the best R(Z_H) estimator during this extreme rainstorm event.