气象学报  2012, Vol. 70 Issue (1): 91-100 PDF
http://dx.doi.org/10.11676/qxxb2012.008

0

#### 文章信息

ZHANG Yu, CHEN Dehui, XUE Jishan, MA Xulin. 2012.

Study of the influence of the humidity factor on estimation in the adaptive observation sensitive region

Acta Meteorologica Sinica, 70(1): 91-100.
http://dx.doi.org/10.11676/qxxb2012.008

### 文章历史

1. 中国气象科学研究院，北京，100081;
2. 南京信息工程大学，南京，210044;
3. 国家气象中心，北京，100081

Study of the influence of the humidity factor on estimation in the adaptive observation sensitive region
ZHANG Yu1,2, CHEN Dehui3 , XUE Jishan1, MA Xulin1
1. Chinese Academy of Meteorological Sciences, Beijing100081, China;
2. Nanjing University of Information Science and Technology, Nanjing210044, China;
3. National Meteorological Centre, Beijing100081, China
Abstract: The ETKF (Ensemble Transform Kalman Filter)-DRY adaptive observation scheme can obtain the relative accurate sensitive region, but it hasn’t contained the humidity factor. This article is based on the ETKF-DRY scheme developed by Ma (2008). According to the difference of the data and the metrics we constructed three different schemes, and compared with the old one. We choose the freezing disaster occurred in the early 2008 as an example. The results show that the sensitive area which is calculated by the old one is more scattered, and there are false sensitive areas when the verification time is equal to the adaptive observation time; with the adding of the humidity factor and the changing of the metrics the false sensitive areas have an apparent decrease; the ETKF-WET-E6 system which only contains the humidity factor has the best results, say, the sensitive areas are more concentrated at the adaptive observation time and there are no false sensitive areas when the verification time is equal to the adaptive observation time.
Key words: Adaptive observation     Humidity factor     Sensitive region     Freezing rain disaster
1 引 言

P(ti+M| H i)表示在ti+M时刻，在常规观测中加入适应性观测资料所得预报误差协方差，由卡尔曼滤波理论(Welch et al，2006)可得
S(ti+M| H i)=Kk P(ti+M| H r)，为加入适应性观测资料后误差协方差的减少量，称作信号方差。其中

 图 1 适应性观测敏感区估算方案流程Fig. 1 Flow chart for the adaptive observation scheme

4 个例试验

 图 2 适应性观测策略简要流程(ti为集合预报初始时刻；ti+m为目标观测时刻，m=1，2…，表示有m个目标观测时刻；tv表示验证时刻)Fig. 2 Flow chart of the decision-making process of the adaptive observation(ti is the initial time of the ensemble prediction; ti+m is the adaptive observation time with m=1，2… meaning the number of the observation; tv is the verification time)

2008年年初的冰冻过程，在1月25—29日雨雪灾害强度达到最大。25日00时(世界时，下同)启动了重大气象灾害预警应急预案。据此选取试验验证时刻tv为2008年1月27日12时，集合预报初始时刻ti为2008年1月25日00时，目标观测时刻ti+m分别为：2008年1月27日12时-48 h(验证时刻前48 h，下同)、-36 h、-24 h、-12 h、-0 h。 4.2 验证区的选取

 图 3 模拟试验确定验证区方案对比Fig. 3 The comparison of the two simulation experiments

 图 4 模拟-2相对模拟-1的850 hPa高度场偏差(彩色区，黑圆点所围区域为偏差的大值区)及500 hPa高度场(实线，单位：gpm)Fig. 4 Differences in the height field at 850 hPa between simulation 1 and simulation 2(the area circled by the dotted line is the large difference area; the black line is the height at 500 hPa)
5 试验结果 5.1 试验资料说明

 图 5 4个不同方案计算所得的目标观测时刻2008年1月27日12时-0 h信号方差(ａ．ETKF-DRY，ｂ．ETKF-WET-E6，ｃ．ETKF-WET-E4，ｄ．ETKF-WET-E5；实线为500 hPa高度场，单位：gpm；彩色区为信号方差大值区，黑色方框区域为验证区)Fig. 5 Signal variance charts at 12:00 UTC 27 Jan 2008 which are calculated by the four schemes(a. ETKF-DRY，b. ETKF-WET-E6，c. ETKF-WET-E4，d. ETKF-WET-E5; the black line is the height at 500 hPa; the colour area is the high signal variance area; the rectangular area is the verification area)
 图 6 ETKF-DRY方案在不同目标观测时刻计算所得的信号方差(a、b、c、d．目标观测时刻为2008年1月27日12时-48 h、-36 h、-24 h、-12 h；实线为500 hPa高度场，单位：gpm；彩色区为信号方差大值区，黑色方框区域为验证区)Fig. 6 Signal variance charts at the different adaptive observation time against the targeting time of 12:00 UTC 27 Jan 2008 which are calculated by the ETKF-DRY scheme(a. -48 h，b. -36 h，c. -24 h，d. -12 h; see the text for farther details; the black line is the height of 500 hPa，the colour area is the high signal variance area，and the rectangular area is the verification area)
 图 7 ETKF-WET-E4和ETKF-WET-E5方案在不同目标观测时刻计算所得的信号方差(a、b． 2008年1月27日12时-36 h，c、d． 2008年1月27日12时-12 h；实线为500 hPa高度场，单位：gpm；彩色区为信号方差大值区，黑色方框区域为验证区)Fig. 7 Signal variance charts at the different adaptive observation time against the targeting time of 12:00 UTC 27 Jan 2008 which are calculated by the ETKF-WET-E4 and ETKF-WET-E5 scheme(a/b. ETKF-WET-E5/ETKF-WET-E4 at -36 h; c/d. ETKF-WET-E5/ETKF-WET-E4 at -12 h; the black line is the height of 500 hPa，the colour area is the high signal variance area，and the rectangular area is the verification area)
 图 8 ETKF-WET-E6方案在不同目标观测时刻计算所得的信号方差(a、b、c、d. 同图 6，实线为500 hPa高度场，单位：gpm；阴影区为信号方差大值区，黑色方框区域为验证区)Fig. 8 Signal variance charts at the different adaptive observation time which are calculated by the ETKF-WET-E6 scheme(The captions are the same as those in Fig. 6)
 图 9 ETKF-DRY(a、b)和ETKF-WET-E6(c、d)方案在验证时刻前12 h计算所得的扰动场(a、c)和信号方差(b、d)对比(实线为500 hPa高度场，单位：gpm；箭头为500 hPa风矢量，阴影区为大值区，黑色方框区域为验证区)Fig. 9 The perturbation field and the signal variance charts at the different adaptive observation time against the targeting time of 12：00 UTC 27 Jan 2008 which are calculated by the ETKF-DRY and ETKF-WET-E6 schemes(a/c. The perturbation field，and b/d. The signal variance field of the ETKF-DRY/ETKF-WET-E6 is at -12 h(see the text for farther details). The black line is the height of 500 hPa，the arrow is the wind vector of 500 hPa，the left shadow area is the high perturbation area，the right shadow area is the high signal variance area，and the rectangular area is the verification area)

ETKF-DRY方案计算结果表明(图 6)，信号方差大值区随着南支槽前西南气流自西南向东北方向移动，最后到达验证区。这说明能表征验证区预报误差方差减少的信号方差大值区也在随着天气方案变化。但该方案计算所得敏感区过于分散，不利于适应性观测的实施，即敏感区估算效果并不理想。

ETKF-WET-E4、ETKF-WET-E5方案也得到与ETKF-DRY类似结果(图 7)，所求得的敏感区过于分散。因此，敏感区估算效果并不理想。相对而言，ETKF-WET-E5方案在度量标准中加入了扰动湿度项，故在验证时刻前12 h计算所得敏感区较ETKF-WET-E4更为集中，所得效果也较好。ETKF-WET-E6方案估算的敏感区(图 8)也随着槽前的西南气流逐渐向验证区靠近，与前述各个方案相比，其敏感区最为集中，利于目标观测地点的具体选定，该方案的计算结果最为理想。

(1)通过加入湿度场的计算能够使目标时刻即为验证时刻的虚假敏感区减少，

(2)对本个例而言：湿度计算的敏感区效果最优。在各个目标时刻所求取的敏感区较另外3个方案更小，并更加利于适应性观测的实施。在验证时刻和观测时刻重合时，信号方差大值区能更好地包含于验证区中，且无虚假敏感区。

 陈德辉,沈学顺.2006.新一代数值预报系统GRAPES研究进展.应用气象学报,17(6):773-777 雷荔傈,谈哲敏.2008.适应性观测及其策略问题.气象科学,28(1):109-118 马旭林.2008.基于集合卡尔曼变换(ETKF)理论的适应性观测研究与应用[D].南京:南京信息工程大学,126 pp 穆穆,王洪利,周菲凡.2007.条件非线性最优扰动方法在适应性观测研究中的初步应用.大气科学,31(6):1102-1112 谭晓伟.2009.CNOP新算法研究及其在目标观测中的应用检验[D].北京:中国科学院大气物理研究所,99 pp 田伟红,庄世宇.2007.适应性观测与集合变换卡尔曼滤波方法介绍.热带气象学报,23(2):201-204 薛纪善, 陈德辉.2008.数值预报系统GRAPES的科学设计与应用.北京:科学出版社,383 pp Baker N L, Daley R. 2000. Observation and background adjoint sensitivity in the adaptive observation-targeting problem. Quart J Roy Meteor Soc, 126: 1431-1454 Bergot T, Hello G, Joly A, et al. 1999. Adaptive observations: A feasibility study.Mon Wea Rev, 127: 743-765 Bishop C H, Toth Z. 1999. Ensemble transformation and adaptive observations. Atmos Sci, 56: 1748-1765 Bishop C H, Etherton B J, Majumdar S J. 2001. Adaptive sampling with the ensemble transform Kalman filter Part Ⅰ: Theoretical aspects. Mon Wea Rev, 129: 420-436 Bouttier F, Courtier P. 2002. Data assimilation concepts and methods. Meteor Training Course Lecture Series, ECMWF Kalman R E. 1960. A new approach to linear filtering and prediction problems. J Basic Engineering, 82: 35-45 Majumdar S J, Bishop C H, Etherton B J. 2002. Adaptive sampling with the ensemble transform Kalman filter Part Ⅱ: Field program implementation. Mon Wea Rev, 130:1356-1369 Pu Z X, Kalnay E, Sela J, et al. 1997. Sensitivity of forecast errors to initial conditions with a quasi-inverse linear method.Mon Wea Rev, 125(10): 2479-2503 Rabier F, Klinker E, Coutier P, et al. 1996. Sensitivity of forecast errors to initial conditions.Quart J Roy Meteor Soc, 122:121-150 Snyder C. 1996. Summary of an informal workshop on adaptive observations and FASTEX. Bull Amer Meteor Soc, 77: 953-965 Welch G, Bishop G. 2006. An introduction to the Kanlman fileter. Comput Inf Sci, 7(1): 1-16 Wu C C, Lin P H, Aberson S, et al. 2005. Dropwindsonde observations for typhoon surveillance near the Taiwan region (DOTSTAR): An overview. Bull Amer Meteor Soc, 86: 787-790