高分辨率合成孔径雷达卫星反演风场资料在中国近海风能资源评估中的应用研究

由于海上测风非常昂贵,实地测风资料严重不足,而卫星反演海面风场资料可以有效地弥补这一缺陷,对近海风能资源评估具有至关重要的意义。近年来,欧洲太空局2002年3月发射的Envisat卫星搭载合成孔径雷达（SAR）图像产品在欧洲近海风能资源评估中得到了广泛应用。文中探讨了空间分辨率约1 km×1 km的SAR卫星反演海面风场资料应用于杭州湾近海风能资源评估的技术方法,通过杭州湾海域的实测风速与SAR卫星反演海面风速数据的对比分析发现：（1）14个实测站点中13个站的相对误差小于20%,其中7个站的相对误差小于10%,平均标准差为2.29 m/s;（2）以SAR卫星反演风速数据为基础计算的风能参数（形状参数和尺度参数）与实测数据计算结果的一致性较高;（3）将该数据同化进入WRF数值模式中,与控制试验相比,大部分检验站点的风速相关系数明显提高,标准差和相对误差也得到改善。SAR卫星反演风场资料可用于中国的近海风能资源评估。     

Kinematic features of a bow echo in southern China observed with doppler radar

A bow echo is a type of mesoscale convective phenomenon that often induces extreme weather and appears with strong reflectivity on radar images. A strong bow echo that developed from a supercell was observed over Foshan City in southern China on 17 April 2011. The intense gusty winds and showers caused huge losses of property and severely affected human lives. This paper presents an analysis of this strong meso- n-scale convective system based on Doppler radar observations. The isolated bow echo exhibited a horizontal scale of about 80 km in terms of reflectivity above 40 dBZ, and a life span of 8 hours. The system originated from the merging of a couple of weakly organized cells in a shear line, and developed into an arch shape as it moved through the shear zone. Sufficient surface moisture supply ensured the convective instability and development of the bow echo. The low-altitude winds retrieved from single Doppler radar observations showed an obvious rear-inflow jet along the notch area. Different from the conventional definition, no book- end anticyclone was observed throughout the life cycle. Very strong slantwise updrafts and downdrafts were recognizable from the retrieved winds, even though the spatial scale of the bow echo was small. Strong winds and induced damage on the surface are considered to have been caused by the mid-level rear-inflow jet and intense convective downdrafts.     

多种单多普勒雷达风场反演方法对比试验

采用典型风场资料对 4种单多普勒天气雷达风场反演方法进行了对比试验 ,这4种方法是 VAP方法 ,涡度 -散度法 ,简单伴随函数方法 ,动力方程组求解方法。试验结果表明 ,不同风场应采用不同反演方法。     

单部多卜勒天气雷达探测低空风切变方法

利用单多卜勒天气雷达风场资料对一维径向、一维方位、二维复合风切变值进行了计算 ,并对机场附近飞机所受气流速度变化进行分析。所得结果旨在定量描述机场附近低空风切变区 ,检验对下击暴流、中气旋、阵风锋探测的结果 ,确定飞机起降潜在危险区。     

台风“碧利斯”的雷达反演及改进

利用两步变分反演方法,研究2006年第4号台风“碧利斯”,对照单部雷达资料和两部雷达资料的反演结果。结果表明：利用单部雷达资料不能很好的反演出台风的风场结构特征,而两部雷达资料在第一步反演背景场时,可以看出台风的涡旋特征。通过检验发现,此次台风过程的水平切向风在最大风速半径内部与Rankine模型相近,外部与Chen 3模型接近。为了更有效地利用单部雷达的观测信息准确定位台风的移动路径,故在单部雷达资料第一步反演背景场时加入模型弱约束。反演结果表明,加入模型弱约束后,反演的背景风场就能看出台风的涡旋特征,通过第二步反演后与两部雷达资料结果相近;加模型弱约束的单部雷达反演结果与两部雷达反演结果间u、v分量的相关系数均在0.8或以上,2-7 km高度层u、v分量的均方根误差明显减小。     

台风“碧利斯”的雷达反演与改进

利用两步变分反演方法研究2006年第4号台风＂碧利斯＂,比较单部雷达资料和双部雷达资料的反演结果。结果表明：利用单部雷达资料不能很好地反演出台风的风场结构特征,而双部雷达资料在第一步反演背景场时,可以显示出台风的涡旋特征。通过检验发现,此次台风过程的水平切向风在最大风速半径内部与Rankine模型相近,外部与Chen 3模型接近。为了更有效地利用单部雷达的观测信息准确定位台风的移动路径,故在单部雷达资料第一步反演背景场时加入模型弱约束。反演结果表明,加入模型弱约束后,反演的背景风场可以看出台风的涡旋特征,通过第二步反演后与双部雷达资料结果相近;加模型弱约束的单部雷达反演结果与双部雷达反演结果间u、v分量的相关系数均为0.8或以上,2—7 km高度层u、v分量的均方根误差明显减小。     

华南一次强飑线过程的三维变分风场反演效果分析

利用三维变分方法对2014年3月30—31日华南一次强飑线过程进行风场反演,经与风廓线雷达探测结果、双多普勒天气雷达反演结果、原始径向速度数据等对比分析,得到如下结论：三维变分方法反演的中低层水平风场与风廓线雷达探测到的结果较为一致,且能很好地表现飑线过境时的风向切变;通过与双多普勒雷达风场反演结果对比发现,两种方法得到的风场空间分布十分相似,均能很好地表现2 km高度上系统内部强带状回波前缘的辐合线以及5 km高度上较弱的辐散;三维变分方法反演的水平风场与径向速度场有较好的一致性,2 km高度强回波带前缘阵风锋处的辐合线位置以及5 km和8 km高度上辐散区的位置均与径向速度场十分吻合;三维变分方法反演的垂直速度能较好地反映该飑线过程中气流的上升和下沉运动,平行于飑线方向的气流变化较小,而系统气流变化主要沿垂直于飑线的方向。三维变分方法反演的飑线系统的三维风场结构合理,反演结果可靠。     

一次消（减）雨试验过程风场特征及作业效果分析

根据济南齐河多普勒雷达体扫资料,采用EVAP方法对山东2009年9月19—20日一次积层混合云降水进行风场反演,反演风场能细致地揭示本次降水过程中尺度风场的演变过程,清楚地给出了风场中尺度结构、辐合线的位置及二者演变特征。各个层次风场变化与常规探空观测的结果基本一致,降水过程前半段存在中尺度辐合线,后期随着冷空气加强南侵并完全控制了中低层,对应的地面降水也由强减弱并逐渐结束。由于反演风场能实时反映出中尺度辐合线的位置,有利于人工影响天气作业的时机及部位的选择。     

多普勒雷达风廓线产品应用初探

利用雷达投入运用一年来的风廓线产品资料，对给桂林站造成大雨以上或大风、冰雹等强对流天气的过程进行分类，挑选几个典型个例，结合多普勒雷达风廓线产品进行研究。得出各类天气形式下风廓线产品的变化特点。     

多普勒雷达风廓线产品应用初探

利用雷达投入运用一年来的风廓线产品资料,对给桂林站造成大雨以上或大风、冰雹等强对流天气的过程进行分类,挑选几个典型个例,结合多普勒雷达风廓线产品进行研究。得出各类天气形式下风廓线产品的变化特点。     

多普勒雷达与L波段雷达垂直风廓线对比分析北大核心

利用2012-2017年阜阳多普勒雷达与L波段雷达测风数据进行对比分析,统计两者的相关性和测量误差,进一步了解多普勒雷达风廓线产品的准确性和可信度。结果表明:两者测风结果一致性较好,风向和风速相关系数分别为0.97和0.94,标准差分别为19.5°和2.65 m·s^(-1)。多普勒雷达风速总体上在同一高度比L波段雷达风速偏小,两者风速相对偏差平均为24.48%;风速标准差随高度增高呈增大趋势,在降水期间对比差值小于非降水;风向标准差在7 km以下呈递减趋势,8 km以上有小幅增加趋势;风速相关系数随高度增加呈增大趋势,除低空偏低以外,其他高度相关系数均较高。     

单多普勒雷达径向风同化的改进

基于现有的单多普勒雷达径向速度资料直接同化和间接同化方法,提出了一种改进同化方法,通过引入反演方程组作为观测算子,将分析单元内径向风的空间变化信息作为同化变量。同时,根据模式的分辨率和精度要求调整分析单元大小,从而能够有效地弥补数据稀释化所带来的高分辨率数据遗失问题。相对于直接同化只同化雷达观测径向风,改进方法不仅包含径向风信息,而且包含径向风的空间变化信息;相对于传统的间接同化,改进方法不直接使用强约束求解析风场,避免了将反演风场完全等同于"观测风场",能够较大程度地克服反演误差对同化效果的影响。通过理想实验和观测系统模拟试验表明:改进方法对于传统的直接同化和间接同化具有一定的兼容性,应用灵活,具有积极的改进意义。     

登陆台风边界层风廓线特征的地基雷达观测

为了分析登陆台风边界层风廓线特征,利用2004—2013年中国东南沿海新一代多普勒天气雷达收集的17个登陆台风资料,采用飓风速度体积分析方法,反演登陆台风的边界层风场结构特征。与探空观测对比表明,利用雷达径向风场可以准确地反演登陆台风的边界层风场结构,其风速误差小于2 m/s,风向误差小于5°。所有登陆台风合成的边界层风廓线显示,在近地层(100 m)以上,边界层风廓线存在类似急流的最大切向风,其高度均在1 km以上,显著高于大西洋观测到的飓风边界层急流高度(低于1 km)。陆地边界层内低层入流强度也明显大于过去海上观测,这主要是由陆地上摩擦增大引起。越靠近台风中心,边界层风廓线离散度越大,其中,径向风廓线比全风速以及切向风廓线离散度更大。将风廓线相对台风移动方向分为4个象限,分析边界层风廓线非对称特征显示,台风移动前侧入流层明显高于移动后侧。最大切向风位于台风移动左后侧,而台风右后侧没有显著的急流特征,与过去理想模拟的海陆差异导致的台风非对称分布特征一致。     

RETRIEVAL OF HORIZONTAL WIND PERTURBATION FIELDS FROM SIMULATED SINGLE DOPPLER RADAR OBSERVATIONS*

The application of the single Doppler radar dataset analysis is usually confined to the assumption that the actual wind is linearly distributed or uniform locally.Following some dynamic features of convective weather,a conceptual model of moderate complexity is constructed,wherewith a horizontal wind perturbation field is retrieved directly from the single Doppler radar measurements.The numerical experiments are based on a 3-D cloud model-generated convective cell,whose radial velocity component is taken as the radar observations that are put into the closed equations based on the conceptual model to retrieve the horizontal wind perturbation field.After the initial field is properly treated,the retrieval equation is solved in terms of the 2-D FFT technique and the sensitivity to noise is examined.Finally,contrast analysis is done of the retrieved and the cloud model output wind fields,indicating the usefulness of the approach proposed in this paper.     

RETRIEVAL OF HORIZONTAL WIND PERTURBATION FIELDS FROM SIMULATED SINGLE DOPPLER RADAR OBSERVATIONS

The application of the single Doppler radar dataset analysis is usually confined to the assumption that the actualwind is linearly distributed or uniform locally.Following some dynamic features of convective weather,a conceptualmodel of moderate complexity is constructed,wherewith a horizontal wind perturbation field is retrieved directly fromthe single Doppler radar measurements.The numerical experiments are based on a 3-D cloud model-generatedconvective cell,whose radial velocity component is taken as the radar observations that are put into the closed equationsbased on the conceptual model to retrieve the horizontal wind perturbation field.After the initial field is properlytreated,the retrieval equation is solved in terms of the 2-D FFT technique and the sensitivity to noise is examined.Finally,contrast analysis is done of the retrieved and the cloud model output wind fields,indicating the usefulness of theapproach proposed in this paper.     

An Integrating VAP Method for Single-Doppler Radar Wind Retrieval

Some traditional methods, such as the velocity-azimuth display (VAD) and the velocity-azimuth processing (VAP), have been widely usedto retrieve the 3-D wind field from single-Doppler radar data because of their relative conceptual and practical simplicity. The advantage of VAD is that it is not affected by small-scale perturbations of the radial wind along the azimuth, to which the VAP method is very sensitive. Nevertheless, the spatial resolution of the VAD method is very poor compared to the VAP method. We show, in this study, that these two retrieval methods are actually related with each other and they are two special applications of a retrieval function based on the azimuthal uniform-wind assumption for a given azimuthal interval [θ1, θ2]. When using this retrieval function to retrieve wind fields, the azimuthal interval used in retrieval can be adjusted according to the requirement of smoothness or resolution. The larger (smaller) the azimuthal interval is, the coarser (finer) the horizontal resolution of retrieved wind field is, and the more insensitive (sensitive) the retrieval method is to small-scale perturbations. Because the full information within the azimuthal interval [θ1, θ2],instead of the information at two terminal points only, i.e., azimuths θ1 and θ2, is used to retrieve the wind fields, this method is referred to as the integrating VAP (IVAP) method, wherein the horizontal wind field is retrieved by using the Doppler velocity over the part of circumference, delimited by the given azimuthal interval times the scan radius. By contrast,the VAP method uses only the velocities at two terminal points of the given azimuthal interval. Therefore, the IVAP method has a filtering function, and the filtering rate can be controlled by adjusting the azimuthal interval. The filter such as that used in the pre-processing of the VAP method is no longer necessary for the IVAP method. When the retrieval azimuthal interval is as large as a whole circumference, the IVAP becomes the VAD. On the other hand, if only two neighboring azimuthal data are used, the IVAP becomes the VAP. The frequency response function of IVAP indicates that the IVAP method can filter out shortwaves, and a larger azimuthal interval leads to stronger filtering ability, therefore a smoother retrieved wind field. The shortwave filter function of the IVAP method is tested by an ideal experiment wherein the radar observations are artificially created by a uniform flow superposed with random disturbances. The VAP and IVAP methods are used in wind retrieval, respectively, and give different results for different azimuthal intervals (i.e., 6°, 12°, 24°, and 48°). Because the VAP method is sensitive to small disturbances, the retrieved winds have larger errors for all different azimuthal intervals. However, the retrieved wind by the IVAP method has smaller errors when the azimuthal interval is larger due to its shortwave filter function. An experiment for an idealized linear wind field is also carried out to demonstrate the effect of the retrieval azimuthal interval on the IVAP method. The results show that a short interval gives the retrieval close to the “true" wind field with a linear distribution. When increasing the interval, the retrieval is smoothed and can represent only the mean wind field.     

WIND SHEAR IDENTIFICATION WITH THE RETRIEVED WIND OF DOPPLER WEATHER RADAR

Wind shear reflects that the wind field is not uniform, which is one of the primary factors which make the retrieval of the wind field difficult. Based on volume velocity process(VVP) wind field retrieval technique, the intensity of wind shear is identified in this paper. After analyzing the traditional techniques that rely on the difference of radial velocity to identify wind shear, a fixed difference among radial velocities that may cause false identification in a uniform wind field was found. Because of the non-uniformity in wind shear areas, the difference of retrieved results between surrounding analysis volumes can be used as a measurement to show how strong the wind shear is. According to the analysis of a severe convective weather process that occurred in Guangzhou, it can be found that the areas of wind shear appeared with the strength significantly larger than in other regions and the magnitude generally larger than4.5 m/(s·km). Besides, by comparing the variation of wind shear strength during the convection, it can be found that new cells will be more likely to generate when the strength is above 3.0 m/(s·km). Therefore, the analysis of strong wind shear's movement and development is helpful to forecasting severe convections.     

单多普勒雷达反演热带气旋近中心风场的VAP扩展应用方法

在尝试运用VAP方法和扩展VAP(EVAP)方法,对"韦帕"台风期间移动天气雷达(CINRAD/CCJ)观测基数据进行单雷达风场反演时,发现部分区域出现风向、风速不合理的情况.分析认为,在热带气旋近中心区域风向风速变化均较剧烈,而上述两种算法的假定条件难以准确代表风场的实际变化.根据热带气旋发展成熟阶段的气旋环流近似轴对称的特点,文中提出一种新的假定条件,即在与热带气旋中心等距离的圆周上局地风速不变、风向均匀变化,对VAP方法进行扩展应用,称为EVAPTC(EVAP for Tropical Cyclone)方法.对理想模型流场试验及对实际个例的反演结果表明,EVAPTC方法简便,反演误差相对较小,能够基本反映出台风近中心的环流特点,尤其是螺旋云带上的风场特征.例如"韦帕"台风螺旋云带内侧有明显的辐合气流进入螺旋云带,螺旋云带尾部的外围气流绕着云体移动,螺旋云带上的强回波区存在着风场扰动等.     

多普勒天气雷达识别低空风切变算法的改进研究

针对CINRAD-CD天气雷达切变线识别产品的局限性,通过对探测资料采取中值滤波、滑动平均等预处理方法,有效填充缺测区探测资料。随后,再对探测资料采用最小二乘拟合法计算得到风场径向速度的径向切变、方位切变及垂直切变,进一步得到风场径向和方位的组合切变,从而得到垂直切变图。通过两次银川降雨过程实例应用,证明上述方法能够较准确地辨识风切变线,可为临近预报提供更科学的参考信息。