INVESTIGATION ON EFFECTS OF INITIAL SCHEMES FOR BINARY TYPHOONS ROKE AND SONCA IN 2011

Based on the Tropical Region Atmospheric Modeling System for South China Sea (TRAMS), Typhoon Roke (1115) and Sonca (1116) in 2011 which have large forecast errors in numerical operation prediction, have been selected for research focusing on the initial scheme and its influence on forecast. The purpose is to find a clue for model improvement and enhance the performance of the typhoon model. Several initialization schemes have been designed and the corresponding experiments have been done for Typhoon Roke and Sonca. The results show that the forecast error of both typhoons’ track and intensity are less using the initial scheme of relocation and bogus just for the weak Typhoon Sonca, compared with using the scheme for both typhoons. By analysis the influence of the scheme on weak typhoon vortex circulation may be the reason that leads to the improvement. All weak typhoons in 2011 to 2012 are selected for tests. It comes to the conclusion that the initial scheme of relocation and bogus can reduce the error of track and intensity forecast. Besides, the height of cloud top in typhoon vortex constructed by bogus is too high according to weak typhoon. It is feasible to develop a bogus which is suitable for weak typhoon.     

华南地区中尺度模式预报的初值影响分析

基于GRAPES 3D-Vat同化系统，介绍了根据静力平衡关系由高度场的增量来计算温度场的增量的方法，并结合常规探空资料同化，针对影响华南地区的一次冷空气过程的个例，分析了常规观测资料对初始场的影响。采用GRAPES区域中尺度模式进行了控制试验和同化试验，观测资料的检验表明，资料同化对GRAPES中尺度模式预报有正面影响，初始时刻预报敏感性较大并且变化较快，靠近地面的低层预报敏感性较为复杂，常规探空资料对高层大气预报能力的提高相对于中低层较小。     

雷达反演水汽在华南前汛期短时临近降水预报应用试验

为了改善数值预报模式短时临近降水预报能力,采用雷达水汽反演同化方案来提高模式初始水汽场质量。以华南精细预报模式对2013年5月进行了冷启动、热启动和热启动加雷达水汽同化等3个对比试验,并进一步详细分析2013年5月8—9日广东西南部到珠三角地区强降雨过程,探讨雷达反演水汽对短时临近降水预报的影响作用。分析结果显示3个试验结果反映精细模式冷启动不利于1~6 h的降水预报,模式热启动能较好预报前6 h的降水,但对随后预报效果不佳。在热启动的基础上,考虑同化雷达反演的水汽后,不仅较好模拟前6 h的降水,也可以预报出6~12 h的降雨带变化趋势。     

川滇黔桂地区下、中三叠统层序划分、对比及层序地层格架

在对研究区不同相带内典型剖面详细观测、层序界面识别基础上,将区内下、中三叠统划分出六个三级层序,分别命名为S₁ ～S₆,其中下三叠统印度阶一个 (S₁ )、奥伦尼阶两个 (S₂ ～S₃),中三叠统安尼阶两个 (S₄～S₅)、拉丁阶一个 (S₆),从年代地层考虑早三叠延续时限为 3.9Ma,则S1 、S2 、S3层序延续的时限平均为 1.3Ma。中三叠统延续的时限为 6.1Ma,则S4、S5、S6 层序所延续的时限平均分别为 2.0 3Ma。按层序底面性质可分出 4个Ⅰ型层序 (S₂ 、S₃、S₄、S₅)和 2个Ⅱ型层序 (S₁ 、S₆),并进行了区域层序对比,进而建立了区内下、中三叠统层序地层格架模型,详细讨论了各层序的次级组成单元 (体系域 )的几何形态和相互关系随时间的迁移所发生的相应变化。     

海滩碎波带波能振荡的多尺度分析

对不同波况条件下实测的一岬间海滩碎波带两个测站的波面数据 ,通过小波变换方法计算了小波能谱 ,并作了显著性检验。结果表明 :( 1 )碎波带波能存在典型的多尺度振荡特性 ;( 2 )在正常波况下 ,波能主要分布于风浪频域 ,入射波自外测站向内测站传播过程中 ,能量具有向低频方向转移的趋势 ;( 3)在高能波况下 ,能量均向长重力波频带增加 ,在长重力波显著尺度 ,二测站之间存在良好的位相一致关系     

IMPROVED FORECASTING OF COLD AIR OUTBREAKS OVER SOUTHERN CHINA THROUGH OROGRAPHIC GRAVITY WAVE DRAG PARAMETERIZATION

The parameterization of gravity wave drag induced by sub-grid scale orography(GWDO), which has been used in the regional model based on the Global/Regional Assimilation and Prediction System for Tropical Mesoscale Model(GRAPES_TMM), is extended to include the effect of mountain flow blocking drag(MBD). The extended scheme is evaluated against non-GWDO parameterization, including a cold air outbreak over southern China and a monthly verification in February 2012. The experiment results show that the GWDO and MBD parameterization both improves the forecasting of the cold air outbreaks over southern China, as well as alleviations of system bias of GRAPES_TMM.(1) The extended scheme alleviates the strong southerly wind and high surface temperature simulation during the cold air outbreak, especially over northern Guangxi and Guangdong(NGG) province, where local high surface temperature simulation reduces nearly 5 degree.(2) The MBD parameterization improves southerly wind simulations over NGG, as well as surface temperature forecasts improvement over Guangxi, Guizhou province and southern Yunnan-Guizhou plateau(YUP), and low level southerly wind simulation improvement over intertidal zone over south China.(3) The formation of MBD is mainly in the mountain area(Wuyi, Daba mountain, east of YUP) and coastal area. The MBD over plateau, which is mainly formed at the west of 105°E, is stronger and thicker than that over Nanling mountain.(4) The improvement of GWDO and MBD parameterization is stable in model physics. MBD parameterization demonstrates more overall improvements in the forecasts than GWDO, and the larger of the model forecast error is, the greater improvements of MBD contribute to. Overall, the extended GWDO scheme successfully improves the simulations of meteorological elements forecasting during cold air outbreaks.     

GRAPES_GFS中三维参考大气的研究:理论设计和理想试验

参考大气的选取对于半隐式半拉格朗日（Semi-Implicit Semi-Lagrangian,简称SISL）模式动力框架的计算精度至关重要。中国气象局数值预报中心自主研发的GRAPES_GFS（Global Regional Assimilation and PrEdiction System,Global Forecast System）采用基于等温大气构造的一维参考大气,该方法求解简单、易于实现,但无量纲气压和位势温度扰动量的数量级较大,降低空间计算精度的同时,由于非线性项较大,使得时间计算精度较低。借鉴近年来世界上各主要业务中心的数值模式框架搭建方法,拟在GRAPES_GFS的动力框架中引入不随时间变化且满足静力平衡的三维参考大气,使得积分过程中参考大气可以尽量地靠近模式大气,提高空间计算精度的同时,减小非线性项的数量级,进而提高时间积分的计算精度。本研究重新推导了引入三维参考大气之后模式动力学方程组的求解过程,通过若干个理想试验验证了理论方法以及代码实现的正确性,说明新的三维参考大气可以有效地提高模式动力框架的计算精度。      

高分辨地形对华南区域GRAPES模式地面要素预报影响的研究

华南区域GRAPES模式动力框架的更新使得高分辨地形数据能够进入模式。引入SRTM数据实现静态数据更新,结合模式内置数据,进行了批量模拟试验;通过站点检验方式,对批量试验结果进行对比,得出以下结论:对比业务使用的Topo10 m地形、Topo30 s地形、SRTM地形和基于SRTM多种插值方案得到的地形,海拔偏差的空间分布和分位数统计都有明显的改善,复杂地形区域的改善效果更显著。通过地面要素平均绝对误差(MAE)箱须图统计和模式西部站点绝对误差(AE)时间序列图对比分析,发现高分辨地形试验的2 m气温和10 m风速MAE和AE有大幅度的改善。高分辨地形对模式静态数据的改善是2 m气温和10 m风速MAE下降的主要原因,地形复杂区域对MAE改善的贡献高于模式其他区域。高分辨地形进入模式后会引起动力过程计算的虚假扰动,适当的滤波平滑能够抑制扰动,从而进一步提高预报精度。      

有云条件下HIRS/3资料的同化及其对“珍珠”台风的影响

卫星红外辐射资料汽接同化容易受到云的影响,大量卫星资料在有云情况下无法直接同化,其原因之一就是初始云结构难以获取.而云结构是同化系统观测算子的重要输入参数,缺乏云的结构信息,观测算子尤法模拟有云环境下的亮温.热带地区以及台风、暴雨等极端天气系统与云密不可分,要改进卫星辐射资料特别是红外资料直接同化的效果,需要获得较高精度的云三维结构.在GRAPES-3DVAR(Global and Regional Assimilation and Prediction Enhanced System,全球/区域同化预报系统)基础上,结合WRF(Weather Research and Forecasting Model,天气研究和预报模式)对"珍珠"台风云结构的模拟,增加模拟的云参数作为GRAPES-3DVAR同化系统中RTTOV模式(fast radiative transfer model for TOVS,快速辐射传输模式)的输入参数,改进了有云环境下卫星资料直接同化的观测算子,研究了红外辐射资料直接同化的方法,并分析其对"珍珠"台风预报的影响.针对NOAAl6(National Oceanic and Atmospheric Administration,美国国家海洋和大气管理局)的HIRS/3(High Resolution Infrared Radiation Sounder,高分辨率红外辐射探测仪3型)第4-8、10、13 19通道辐射亮温的同化试验表明,结合初始云结构信息改进同化系统观测算子,可以将红外辎射资料的观测信息同化到模式之中,改进台风温度、湿度结构,对短时降水预报有正面影响.     

MODIS BRIGHTNESS TEMPERATURE DATA ASSIMILATION UNDER CLOUDY CONDITIONS: METHODS AND IDEAL TESTS

Clouds have important effects on the infrared radiances transmission in that the inclusion of cloud effects in data assimilation can not only improve the quality of the assimilated atmospheric parameters greatly, but also minimize the initial error of cloud parameters by adjusting part of the infrared radiances data. On the basis of the Grapes-3D-var (Global and Regional Assimilation and Prediction Enhanced System), cloud liquid water, cloud ice water and cloud cover are added as the governing variables in the assimilation. Under the conditions of clear sky, partly cloudy cover and totally cloudy cover, the brightness temperature of 16 MODIS channels are assimilated respectively in ideal tests. Results show that when the simulated background brightness temperatures are lower than the observation, the analyzed field will increase the simulated brightness temperature by increasing its temperature and reducing its moisture, cloud liquid water, cloud ice water, and cloud cover. The simulated brightness temperature can be reduced if adjustment is made in the contrary direction. The adjustment of the temperature and specific humidity under the clear sky conditions conforms well to the design of MODIS channels, but it is weakened for levels under cloud layers. The ideal tests demonstrate that by simultaneously adding both cloud parameters and atmospheric parameters as governing variables during the assimilation of infrared radiances, both the cloud parameters and atmospheric parameters can be adjusted using the observed infrared radiances and conventional meteorological elements to make full use of the infrared observations.