GRAPES_3Dvar中雷达径向风同化改进观测算子的应用

Application of Advanced Observation Operator of Doppler Radar Radial Velocity Assimilation in GRAPES_3Dvar

在GRAPES_3Dvar系统中添加雷达径向风预处理模块以及雷达径向风同化模块(包含传统观测算子与改进算子)。对GRAPES_3Dvar系统中雷达同化算子进行正确性检验后,利用该系统对2014年6月6日03时(UTC)北京南郊多普勒雷达径向风资料进行同化试验,分析不同观测算子对于初始场以及降水模拟的影响。研究结果表明,同化雷达径向风资料加强了对流系统低层中小尺度辐合辐散信息,同时综合分析风场的径向分量与切向分量,改进算子对于初始场的调整效果优于传统观测算子。对比分析累计降水模拟,采用改进算子的结果在能够准确模拟出降水雨带的位置和量级的同时极大程度地抑制了虚假降水的范围,降水评分也较好地验证了改进观测算子的优越性。

On the basis of GRAPES_3Dvar system, the insertion of pre-processed module and assimilation module for Doppler radar radial velocity assimilation are conducted （including traditional and advanced emulators respectively）. In this paper, the rain case on 6 June 2014 is selected to perform the assimilation and prediction experiments using the GRAPES_3Dvar system after the accuracy test, analyzing the impact on the adjustment of initial conditions and precipitation simulations through different emulators. The results indicate that the mesoscale information increases after assimilating radar radial velocity. Moreover, by comprehensively analyzing the radial and tangential components, the results using the advanced emulator approach are closer to observation and reference wind fields compared with those using the traditional one. In addition, the accumulated precipitation result adopting the advanced emulator can better simulate the location and magnitude of precipitation and effectively restrain the virtual rainfall among all tests, the test scores also illustrate this conclusion.