单多普勒雷达二维风场反演——Extended VPP方法

EXTENDED VPP (EVPP) METHOD FOR RETRIEVAL OF HORIZONTAL WIND FROM SINGLE DOPPLER RADAR DATA

文中提出的EVPP(Extended Velocity Plan Processing)方法是根据风场连续及局部分片光滑的性质,假定在一很小的分析窗内风矢量保持线性变化关系,使得在分析窗内反演得到的风场在矢径上的投影与实际观测的多普勒雷达径向速度场之间的误差达到最小,并利用零速度线这一特点对问题的求解做出约束,使得所得解尽可能达到全局最优,即:(1)利用变分的思想求出水平风速的大小及方向;(2)利用风场局部分片光滑,对风的方向做出矫正。虽然由于风场受到线性假设这一限制,使得其使用的范围得到了局限,但是在本文中利用零速度线这一特征使得反演的结果得到了进一步的提高。文中利用模拟的理想个例和实际观测到的个例分别进行了研究分析。为了证明其理论的可行性,首先我们利用三维冰雹云模式模拟了辽宁2005年7月8日的一次冰雹云过程,考虑到个例的代表性,我们选择了辐合、辐散、旋转以及均匀风场等4种类型的流场结构进行了模拟分析,同时为了验证该种方法对于资料质量的依赖性以及噪声的敏感性,我们在理想质量中加入了随机噪声。从对模拟风场的反演,证明了该方法理论上的可行性,而且噪声对其结构的影响不大。然后针对台风、飑线、强对流等实际天气过程作了进一步分析。最后对反演结果进行了误差分析,结果表明这一方法能较好地反演出中尺度系统的水平流场结构。

The EVPP （Extended Velocity Plan Processing） method is based on the continuity and pieceswise smooth character of wind field. That wind vectors remain linear in a small analyzing window is assumed to minimize the deviations of the radial projections of retrieved wind vectors from the observed Doppler radial velocities within the window, and the zero-velocity line is used to constrain solving to obtain the global optimum solution. That is （1） ： utilizing variational principle to get the value and direction of horizontal winds, （2） ： using wind field＇s local piecewise smooth within the window to rectify the direction of retrieval winds. Under the assumption that the horizontal wind vectors in a small window are uniform, the horizontal wind components can be derived in real time by single-Doppler radar Velocity Plan Processing （VPP）. In this paper, the VPP method is extended. Although the linear assumption of wind field in the retrieval window limits the EVPP＇s application, the utilization of the important property of zero velocity line in Doppler radar velocity field further improve the retrieval results. To show the feasibility of the theory on which the EVPP method is based, the hail-storm on 8 July 2005 in Liaoning was simulated using the 3-D hail cloud model of IAP. In the simulation process, we first get the wind field by the model which is regarded as the true wind field in the atmosphere. Then we project the wind field on the radar dirction and this result is regarded as the observation radial velocity field from the Doppler radar. The model-generated Doppler data used in the retrieval of the hail-storm wind field by the EVPP method. The comparison of the simulation and retrieval on the four typical flow fields of convergent, divergent, rotational, and uniform flow patterns selected from the simulated hailstorm suggest that the theory is feasible. And meantime, in order to examine the sensitivity of the EVPP method to noises and the dependence of the retrieved results show that the noises only have minor influence on the structure of flow field. At last, the accuracy of the retrieval from the real time data including squall line, typhoon, and strong cumulonimbus cloud, was examined against the observation of the dual Doppler radar. The errors analysis and verification of the simulated and real cases indicates that the EVPP method is able to reasonably retrieve and describe the structure of the horizontal flow field of mesoscale systems.