节点文献

单多普勒天气雷达反演二维风场的方法研究

A Retrieval Study of Two-Dimensional Wind Field from Single-Doppler Radar Data

【作者】 王俊

【导师】 吴增茂;

【作者基本信息】 中国海洋大学 , 气象学, 2004, 硕士

【摘要】 我国新一代多普勒天气雷达网建设正在积极进行,山东5部多普勒雷达已建成并投入业务应用,因此,雷达产品的开发应用研究成为当务之急。山东5部雷达的风场反演产品只有VAD产品,雷达资料中包含的风场信息还没有得到充分的开发。因此,利用DOPPLER雷达观测信息,反演中尺度天气系统风场的研究,就成了近年来国内外的热点课题,它将对深入研究中尺度天气系统结构与动力学特征,以及改进天气预报方法具有重要意义。 本文在忽略上升气流而考虑粒子下落速度影响的条件下,导出了不同高度上的水平径向速度计算公式,然后利用VAP方法来反演二维水平风场,这一方法简称为EVAP法。误差分析表明:即使上升速度为2.0m.s-1,仰角为20.0°,忽略上升气流产生的相对误差一般也不超过10%;当仰角为4.5°时,相对误差不超过5%。所以在对流不是很强时,忽略上升气流引起的误差是比较小的,对风场反演结果影响不大。 利用k-邻域频数法对714SDN型多普勒天气雷达原始资料进行预处理,分析了参数的选取对处理结果的影响,并根据应用中出现的问题,对该方法作了补充完善,使补测效果更加有效。主要结果为:利用k-邻域频数法除孤立点时,统计窗口的大小和k1值的选取对结果有重要影响。在统计窗口固定时,应根据原始数据的质量选择合适的k1值,以达到最佳效果。补缺测值时,在统计窗口一定的情况下,存在最大可补测点数,并且这一数目占总数目的比例很低。因此,在补缺测点时,k2应取得小一些,使填补的点数尽可能多,而不必担心影响整体数据结构。另外,在补缺测值时,仅利用xmax对应区间的中值来补缺测值有时效果不明显,进一步利用次最大频率xmax1所对应区间的中值,可以更加有效地填补缺测点。 利用中尺度模式MM5模拟数据试验表明:二维双线性插值对风场结构影响较小,三维插值对结果的影响非常明显,某些区域插值后的结果变小甚至为零。另一方面是气旋性风场插值后强度变弱、范围变小;VAP法反演的二维风场,误差主要集中在切向分量,有时误差是很大的。对径向速度进行平滑滤波后,可以提高切向速度反演的精度。 EVAP法和VAP法反演的暴雨降水个例的风场结果表明:VAP法只能利用最低仰角的PPI资料反演低层的风场,由于风场不在同一等高面上,可能产生一些虚假的系统结构,因此在实际应用中要结合其它资料仔细分析。对于EVAP法,主要的一个缺点是低层风场的范围太小,不利于了解系统低层的整体风场结构。其优点是可以反演不同高度上的二维风场,有助于了解系统的三维结构。 2003年4月17日强降水个例分析表明,EVAP法反演的风场与MMS模拟的风场相比钨第一阶段的降水过程,二者非常一致,Zkm以上高度,均匀乒畅中存在一一条中尺度辐合线,与之对应的是强雷达回波带;而Zkm以下高度风的辐合不明显。第二阶段的降水过程,二者相比差别较大,850Inb模式模拟风场与Zkm.高度雷达反演风场相比,模式乒畅主要是西南风转东南风,而反演风场主要是东南尸呀口静风区,600Inb与4.okm高度的乒畅相比,模式主要是西南风,而反演乒畅为南偏东风转西南风。

【Abstract】 The new era Doppler Radar network of China is building, and 5 sets of Doppler Radar are installed in Shandong province. So, the study is very necessary to exploit and apply radar productions.In the paper, the EVAP (extended VAP method) method is developed, the procedure is firstly to calculate the radial velocity at attitude with the ascend velocity ignored and descent velocity of particle considered, secondly to retrieve Two-Dimensional Wind Field by VAP method. The error analysis shows: when ascent velocity is 2.0 m-s-1 and elevation is 20 degree, the error doesn’t exceed 10 percent It indicates the error of ignoring descent velocity is very little when convection is not very strong.A"K- Neighborhood Frequency" method is used to preprocess original data of Doppler weather radar, and the effect of parameter to the preprocessing result is analysed. The method is modified according to practical need, the main results are: when eliminating isolated point, the result is influenced by the statistical size of area and value of k1. When filling up absence point, there is a maximum number of points filled up by "K- Neighborhood Frequency" method. For the parameter k2, a small value is selected in order to fill up more absence points. However, it is not enough to fill up the absence point only using mid-value of maximum Frequency area, it is better to fill up the absence point also using mid-value of second maximum Frequency area.The test by using model MM5 simulated results shows that, two-dimensional interpolation will produce smaller impact, but three-dimensional interpolation will produce notable effect. The wind error retrieved with VAP method mainly focus on tangential velocity, sometimes the error is very big. In addition, filting the radial velocity can improve the accuracy of tangential velocity retrieved with VAP method.Comparing the wind field retrieved with EVAP and VAP method shows that the VAP method just can use the PPI data of the lowest elevation to retrieve the lower wind field, because the wind is not in the same altitude, it often produce phonily structure, that is the main shortcoming of VAP method. For the EVAP method, the main shortcoming is the range in the low high too small, it is not benefit to realize the whole wind field structure for lower high layer. The advantage is that it can retrieve the two-dimensional wind field in different high layer benefiting to realize the three-dimensional wind field structure of weather system.The rainstorm example of April 17th, 2003 is analysed and the result indicates that comaring to the wind field retrieved by EVAP method and simulated by MM5 model, the first phase the result is very consistent, above 2km there is a meso-scale divergence line ,in the same site the radar echo is higher than 40dbz. Below 2km the meso-scale divergence line is not clear. In the second precipitation phase it is difference between the wind field retrieved by EVAP and simulated by MM5 model.

  • 【分类号】TN959
  • 【被引频次】16
  • 【下载频次】487
节点文献中: 

本文链接的文献网络图示:

本文的引文网络