Assimilating Doppler Radar Data for Storm-Scale Numerical Prediction Using an Ensemble-based Variational Method

三维变分(3DVAR)数据同化(DA)在大多数业务数值天气预报(NWP)中心使用，但主要用于大尺度静力流动。这种技术不能直接推广到对流尺度的非静力流动，因为在这种尺度上不存在简单的平衡关系来连接不同的状态变量，从而得到与模型动力学和物理一致的分析。此外，对流尺度流动的高度空间和时间间歇性使得通常用于大型3DVAR系统的静态背景误差统计无效。在对流尺度上，多普勒雷达是唯一能够提供足够高的空间和时间分辨率的观测以进行动态预报的业务仪器。近年来，集合卡尔曼滤波(EnKF)技术在对流尺度雷达资料同化中显示出巨大的应用前景。然而，该方法并不像变分技术那样成熟，仍然需要进行大量的研究。EnKF的标准实现在计算上也很昂贵。在这项研究中，首席调查员将开发新的有效方法，寻求结合3DVAR和EnKF的优点在对流尺度上应用。具体而言，首席调查员将：(I)继续开发适合对流尺度的3DVAR数据同化策略，并审查适当的方程约束，将三个风分量与热力学场耦合，并同时确定一致的微物理变量；(Ii)将集合得出的背景误差统计应用于3DVAR，以创建一个有效的双分辨率EnKF-3DVAR混合框架，该框架结合了这两种方法的先进特征。首席调查员将使用来自运行中的WSR-88D雷达的数据以及来自美国国家科学基金会大气协作自适应传感工程研究中心的四个雷达试验台的补充数据来评估所产生的技术的性能。知识价值该项目建立在国家科学基金会以前资助的个人投资项目成果的基础上，侧重于一种新的数据同化战略，该战略结合了3DVAR和EnKF的优点，特别适用于对流规模的流动和雷达数据的同化。研究结果可用于为高分辨率风暴尺度数值预报模式提供初始条件，该模式目前正在业务预报中心以较粗的网格间距进行测试。这项研究可能会增进对风暴尺度数据同化和动力学的理解，并导致更好地检测雷暴灾害和改进定量降水预报。更广泛的影响这项研究将有助于从国家对WSR-88D雷达的投资中获得最大利益。它还将加快WSR-88D雷达数据在业务和研究数值预报中的使用。这项研究将通过对研究生和本科生的支持和指导产生教育效益。

Three dimensional variational (3DVAR) data assimilation (DA) is being used at most operational numerical weather prediction (NWP) centers but mainly in the context of large-scale hydrostatic flows. Such a technique cannot be directly extended to convective-scale, non-hydrostatic flows because no simple balance relations exist at such scales to interlink different state variables so as to arrive at analyses that are consistent with model dynamics and physics. Furthermore, the high spatial and temporal intermittency of convective-scale flows renders the static background error statistics typically used in large-scale 3DVAR systems invalid. At the convective scale, Doppler radar is the only operational instrument capable of providing observations of sufficiently high spatial and temporal resolution for dynamic prediction. Recently, the ensemble Kalman filter (EnKF) technique has shown great promise for convective-scale radar data assimilation. The method is, however, not as mature as variational techniques and still requires much research. The standard implementation of EnKF is also computationally expensive. In this research, the Principal Investigator will develop new efficient methods that seek to combine the strengths of 3DVAR and EnKF for application at the convective scale. Specifically, the Principal Investigator will: (i) Continue to develop 3DVAR data assimilation strategies suitable for the convective scale and examine appropriate equation constraints which couple the three wind components with the thermodynamic fields and simultaneously determine consistent microphysical variables; (ii) Apply the ensemble-derived background error statistics to 3DVAR to create an efficient dual-resolution EnKF-3DVAR hybrid framework which incorporates advanced features of both methods. The Principal Investigator will evaluate the performance of the resulting techniques using data from the operational WSR-88D radars as well as complementary data from the four-radar testbed of the NSF Engineering Research Center for Collaborative Adaptive Sensing of the Atmosphere. Intellectual Merit The project builds on the achievements of previous NSF-funded projects of the PIs and focuses on a new data assimilation strategy which combines the advantages of 3DVAR and EnKF and is particularly suitable for convective-scale flows and assimilation of radar data. The research results can be used to provide initial conditions for high-resolution storm-scale NWP models which now are being tested at coarser grid spacing at operational forecast centers. The research will potentially improve the understanding of storm-scale data assimilation and dynamics, and lead to better detection of thunderstorm hazards and improved quantitative precipitation forecasting. Broader Impact The research will help draw maximum benefit from the Nation's investment in the WSR-88D radars. It also will accelerate the use of the WSR-88D radar data in operational and research NWP. This research will produce educational benefits through the support and mentoring of graduate and undergraduate students.