高级检索

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于多重网格策略的NLS-3DVar资料融合方法及其在气温数据融合中的应用

张璐 田向军 刘宣飞 师春香

张璐, 田向军, 刘宣飞, 师春香. 基于多重网格策略的NLS-3DVar资料融合方法及其在气温数据融合中的应用[J]. 气候与环境研究, 2017, 22(3): 271-288. doi: 10.3878/j.issn.1006-9585.2016.16140
引用本文: 张璐, 田向军, 刘宣飞, 师春香. 基于多重网格策略的NLS-3DVar资料融合方法及其在气温数据融合中的应用[J]. 气候与环境研究, 2017, 22(3): 271-288. doi: 10.3878/j.issn.1006-9585.2016.16140
Lu ZHANG, Xiangjun TIAN, Xuanfei LIU, Chunxiang SHI. NLS-3DVar Data Fusion Method Based on Multigrid Implementation Strategy and Its Application in Temperature Data Fusion[J]. Climatic and Environmental Research, 2017, 22(3): 271-288. doi: 10.3878/j.issn.1006-9585.2016.16140
Citation: Lu ZHANG, Xiangjun TIAN, Xuanfei LIU, Chunxiang SHI. NLS-3DVar Data Fusion Method Based on Multigrid Implementation Strategy and Its Application in Temperature Data Fusion[J]. Climatic and Environmental Research, 2017, 22(3): 271-288. doi: 10.3878/j.issn.1006-9585.2016.16140

基于多重网格策略的NLS-3DVar资料融合方法及其在气温数据融合中的应用

doi: 10.3878/j.issn.1006-9585.2016.16140
基金项目: 

公益性行业(气象)科研专项重大项目 GYHY201506002

国家重点研发计划项目 2016YFA0600203

详细信息
    作者简介:

    张璐, 女, 1991年出生, 硕士研究生, 主要从事数据同化研究。E-mail: wisheszhanglu@126.com

    通讯作者:

    刘宣飞, E-mail: liuxf@nuist.edu.cn

  • 中图分类号: P40

NLS-3DVar Data Fusion Method Based on Multigrid Implementation Strategy and Its Application in Temperature Data Fusion

Funds: 

Special Fund for Meteorological Research in the Public Interest GYHY201506002

National Key Research and Development Program of China 2016YFA0600203

  • 摘要: 将多重网格策略引入NLS-3DVar(Non-linear Least Squares-based on Three-dimensional Variational DataAssimilation,非线性最小二乘三维变分同化)方法,进而应用于2400多个国家级气象观测站逐时气温数据和NCEP再分析气温数据的融合,得到中国区域空间分辨率1°×1°,时间分辨率为6小时的气温融合产品。分别从单重网格(分辨率1°×1°)和双重网格(分辨率由2°×2°到1°×1°)利用2014年1~12月(4、5月除外)的独立检验数据考察NLS-3DVar气温融合产品质量,验证基于多重网格策略的NLS-3DVar方法的优越性。在单重网格下,与广泛应用于气象行业的Cressman插值产品(均方根误差和相关系数的年平均值分别为1.961℃ d-1和0.924)相比,NLS-3DVar产品全年始终具有最小的均方根误差和最大的相关系数,年平均值分别为1.915℃ d-1和0.929;站点间误差分析进一步表明,NLS-3DVar产品在大多数检验站点精度更高,在新疆、甘肃、云南、陕西等地区尤为突出;加入双重网格策略的NLS-3DVar产品与单重网格的NLS-3DVar产品误差对比显示,均方根误差年平均值分别为1.649℃ d-1和1.711℃ d-1,相关系数年平均值分别为0.970和0.968,二者在均方根误差和相关系数的表现上都极为相似,即双重网格NLS-3DVar气温产品尽管对观测数据采取了稀疏化处理,但依旧维持了原有的产品精度,并且在计算效率上提高了1倍多。而与同样在双重网格下基于多尺度的STMAS(Space-Time MultiscaleAnalysis System)算法相比,双重网格的NLS-3DVar产品在产品精度上同样占据优势,在计算效率上单位时次耗时与STMAS算法几乎相当。
  • 图  1  (a)中国2400多个国家级气象站分布,(b)中国3万余个自动观测站(包括国家站和区域站)分布

    Figure  1.  Distributions of (a) more than 2400 national automatic weather stations (AWS) and (b) more than 30, 000 national and regional AWS over China

    图  2  历史采样法示意图

    Figure  2.  Sketch map of historical sampling

    图  3  2014年7月13日06时全国气温场(单位:℃)分布:(a)独立检验样本;(b)NLS-3DVar产品;(c)Cressman产品

    Figure  3.  Spatial distributions of temperature over China at 0600 UTC 13 July 2014: (a) The independent validation data; (b) NLS-3DVar (Non-linear Least Squares-based on 3DVar) data; (c) Cressman data

    图  4  2014年(4、5月除外)NLS-3DVar产品、Cressman产品和背景场与独立检验样本的均方根误差(单位:℃ d–1)的逐日变化

    Figure  4.  Daily variations of RMSEs (units: ℃ d–1) between NLS-3DVar product, Cressman product, background field (Xb) and the independent validation data in 2014 except for April and May

    图  5  图 4,但为相关系数的逐日变化

    Figure  5.  As in Fig. 4, but for daily variations of correlation coefficients

    图  6  2014年(4、5月除外)NLS-3DVar产品、Cressman产品、背景场Xb与独立检验样本的月平均RMSE(单位:℃ d–1)和相关系数的Taylor图

    Figure  6.  Taylor diagram of monthly averaged RMSEs (units: ℃ d–1) and correlation coefficients between NLS-3DVar product, Cressman product, background field Xb and the independent validation data in 2014 except for April and May

    图  7  2014年(4、5月除外)NLS-3DVar产品、Cressman产品与独立检验样本的SDV和相关系数的Taylor图

    Figure  7.  Taylor diagram of SDV (standard deviation) and correlation coefficients between NLS-3DVar product, Cressman product and the independent validation data in 2014 except for April and May

    图  8  Cressman产品、NLS-3DVar产品与独立检验样本的RMSE和相关系数的Taylor图(只给出Cressman产品与NLS-3DVar产品的RMSE差值在1.0~1.5℃ d–1范围的站点)

    Figure  8.  Taylor diagram of RMSEs and correlation coefficients between NLS-3DVar product, Cressman product and the independent validation data (only those stations with RMSE differences between Cressman and NLS-3DVar products within 1.0~1.5℃ d–1 are counted)

    图  9  Cressman产品、NLS-3DVar产品与独立检验样本的RMSE差值在1.0~1.5℃ d–1范围的站点分布图

    Figure  9.  Spatial distribution of stations with RMSE differences within the range of 1–1.5℃ d–1 between Cressman and NLS-3DVar products

    图  10  Cressman产品、NLS-3DVar产品与独立检验样本的RMSE和相关系数的Taylor图(只给出Cressman产品与NLS-3DVar产品的RMSE差值 大于1.5℃ d–1的站点)

    Figure  10.  Taylor diagram of RMSEs and correlation coefficients between NLS-3DVar product, Cressman product and the independent validation data (only those stations with RMSE differences between Cressman and NLS-3DVar products greater than 1.5℃ d–1 are counted)

    图  11  Cressman产品、NLS-3DVar产品与独立检验样本的RMSE差值(单位:℃ d–1)分布

    Figure  11.  Spatial distribution of RMSE differences (units: ℃ d–1) between Cressman product and NLS-3DVar product

    图  12  2014年7月2日00时全国气温场(单位:℃)分布:(a)有效独立检验样本;(b)单重网格NLS-3DVar产品;(c)双重网格NLS-3DVar产品;(d)STMAS算法产品

    Figure  12.  Spatial distributions of temperature over China at 0000 UTC 2 July 2014: (a) The independent validation data; (b) single NLS-3DVar product; (c) multigrid NLS-3DVar product; (d) STMAS (Space–Time Multiscale Analysis System) algorithm product

    图  13  2014年10月单重网格NLS-3DVar产品、双重网格NLS-3DVar产品、STMAS算法产品与独立检验样本的(a)均方根误差和(b)相关系数

    Figure  13.  (a) RMSEs and (b) correlation coefficients between products of single NLS-3DVar, multigrid NLS-3DVar, STMAS algorithm product and the independent validation data during October 2014

    图  14  2014年单重网格NLS-3DVar产品、双重网格NLS-3DVar产品、STMAS算法产品与独立检验样本的逐月平均(a)均方根误差和(b)相关系数

    Figure  14.  Monthly mean (a) RMSEs and (b) correlation coefficients between single NLS-3DVar product, multigrid NLS-3DVar product, STMAS algorithm product and the independent validation data in 2014

    图  15  2014年单重网格NLS-3DVar方法、双重网格NLS-3DVar方法和STMAS算法逐月平均单位时次同化过程耗时(单位:s)统计

    Figure  15.  Statistics of monthly mean time-consuming (units: s) of a single-time assimilation by single NLS-3DVar, multigrid NLS-3DVar, and STMAS algorithm in 2014

    表  1  三组试验设计

    Table  1.   Design of three groups of experiments

    网格数目 融合方法1 融合方法2 空间分辨率 时间分辨率
    试验一 单重网格 NLS-3DVar Cressman 1°×1° 6小时
    试验二 单重/双重 单重NLS-3DVar 双重NLS-3DVar 粗网格:2°×2°细网格:1°×1° 6小时
    试验三 双重网格 NLS-3DVar STMAS算法 粗网格:2°×2°细网格:1°×1° 6小时
    下载: 导出CSV

    表  2  2014年(4、5月除外)NLS-3DVar产品、Cressman产品、背景场Xb与独立检验样本的月平均RMSE(单位:℃ d–1

    Table  2.   Monthly averaged RMSEs (units: ℃ d–1) between NLS-3DVar product, Cressman product, background field Xb and the independent validation data in 2014 except April and May

    月份 月平均RMSE/℃ d–1
    NLS-3DVar产品 Cressman产品 Xb
    1月 2.137 2.193 3.107
    2月 1.888 1.931 2.993
    3月 1.967 2.041 2.862
    6月 1.810 1.863 2.731
    7月 1.940 1.982 2.821
    8月 1.862 1.908 2.774
    9月 1.797 1.836 2.652
    10月 1.914 1.955 2.704
    11月 1.812 1.861 2.664
    12月 2.008 2.058 2.956
    下载: 导出CSV
  • [1] Chen D L, Ou T H, Gong L B, et al. 2010. Spatial interpolation of daily precipitation in China:1951-2005[J]. Advances in Atmospheric Sciences, 27(6):1221-1232, doi: 10.1007/s00376-010-9151-y.
    [2] Cressman G P. 1959. An operational objective analysis system[J]. Mon. Wea. Rev., 87(10):367-374, doi:10.1175/1520-0493(1959)087<0367:AOOAS>2.0.CO;2.
    [3] Dennis J E Jr, Schnabel R B. 1996. Numerical Methods for Unconstrained Optimization and Nonlinear Equations (Classics in Applied Mathematics)[M]. Philadelphia:SIAM, 378pp.
    [4] Evensen G. 2004. Sampling strategies and square root analysis schemes for the EnKF[J]. Ocean Dynamics, 54(6):539-560, doi: 10.1007/s10236-004-0099-2.
    [5] 冯锦明, 赵天保, 张英娟. 2004.基于台站降水资料对不同空间内插方法的比较[J].气候与环境研究, 9(2):261-277. doi: 10.3969/j.issn.1006-9585.2004.02.004

    Feng Jinming, Zhao Tianbao, Zhang Yingjuan. 2004. Intercomparison of spatial interpolation based on observed precipitation data[J]. Climatic and Environmental Research (in Chinese), 9(2):261-277, doi:10.3969/j.issn.1006-9585. 2004.02.004.
    [6] Li W, Xie Y F, Deng S M, et al. 2010. Application of the multigrid method to the two-dimensional Doppler radar radial velocity data assimilation[J]. J. Atmos. Oceanic Technol., 27(2):319-332, doi:10.1175/2009JTECHA 1271.1.
    [7] Li W, Xie Y F, He Z J, et al. 2008. Application of the multigrid data assimilation scheme to the China Seas' temperature forecast[J]. J. Atmos. Oceanic Technol., 25(11):2106-2116, doi: 10.1175/2008JTECHO510.1.
    [8] Li W, Xie Y F, Han G J. 2013. A theoretical study of the multigrid three-dimensional variational data assimilation scheme using a simple bilinear interpolation algorithm[J]. Acta Oceanologica Sinica, 32(3):80-87, doi: 10.1007/s13131-013-0292-6.
    [9] Liu C S, Xiao Q N, Wang B. 2009. An ensemble-based four-dimensional variational data assimilation scheme. Part Ⅱ:Observing system simulation experiments with advanced research WRF (ARW)[J]. Mon. Wea. Rev., 137(5):1687-1704, doi: 10.1175/2008MWR2699.1.
    [10] 苗春生, 程远, 王坚红, 等. 2015.中国风云卫星与海洋卫星近海SST资料融合技术及应用研究[J].地球科学进展, 30(10):1127-1143. doi: 10.11867/j.issn.1001-8166.2015.10.1127

    Miao Chunsheng, Cheng Yuan, Wang Jianhong, et al. 2015. Data fusion of offshore SST from China FY and HY2 satellites and its application[J]. Advances in Earth Science (in Chinese), 30(10):1127-1143, doi: 10.11867/j.issn.1001-8166.2015.10.1127.
    [11] 任芝花, 张志富, 孙超, 等. 2015.全国自动气象站实时观测资料三级质量控制系统研制[J].气象, 41(10):1268-1277. doi: 10.7519/j.issn.1000-0526.2015.10.010

    Ren Zhihua, Zhang Zhifu, Sun Chao, et al. 2015. Development of three-step quality control system of real-time observation data from AWS in China[J]. Meteorological Monthly (in Chinese), 41(10):1268-1277, doi: 10.7519/j.issn.1000-0526.2015.10.010.
    [12] 申思. 2015. 降维投影四维变分同化方法的改进及在GRAPES全球模式中的初步应用[D]. 中国科学院大学博士学位论文.

    Shen Si. 2015. Improvements for dimension-reduced projection 4DVar and its preliminary application in GRAPES-GFS[D]. Ph. D. dissertation (in Chinese), University of Chinese Academy of Sciences.
    [13] Taylor K E. 2001. Summarizing multiple aspects of model performance in a single diagram[J]. J. Geophys. Res., 106(D7):7183-7192, doi: 10.1029/2000JA002008.
    [14] Tian X J, Feng X B. 2015. A non-linear least squares enhanced POD-4DVar algorithm for data assimilation[J]. Tellus A, 67:25340, doi: 10.3402/tellusa.v67.25340.
    [15] Tian X J, Feng X B, Zhang H Q, et al. 2016. An enhanced ensemble-based method for computing CNOPs using an efficient localization implementation scheme and a two-step optimization strategy:Formulation and preliminary tests[J]. Quart. J. Roy. Meteor. Soc., 142(695):1007-1016, doi: 10.1002/qj.2703.
    [16] Tian X J, Xie Z H, Sun Q. 2011. A POD-based ensemble four-dimensional variational assimilation method[J]. Tellus A, 63(4):805-816, doi:10. 1111/j.1600-0870.2011.00529.x.
    [17] Wang B, Liu J J, Wang S D, et al. 2010. An economical approach to four-dimensional variational data assimilation[J]. Advances in Atmospheric Sciences, 27(4):715-727, doi: 10.1007/s00376-009-9122-3.
    [18] Xie Y, Koch S, McGinley J, et al. 2011. A space-time multiscale analysis system:A sequential variational analysis approach[J]. Mon. Wea. Rev., 139(4):1224-1240, doi: 10.1175/2010MWR3338.1.
    [19] Xie Y, Koch S E, McGinley J A, et al. 2005. A sequential variational analysis approach for mesoscale data assimilation[C]//Proceedings of the 21st Conference on Weather Analysis and Forecasting/17th Conference on Numerical Weather Prediction. Washington, D. C.:American Meteor Society.
    [20] 熊秋芬, 黄玫, 熊敏诠, 等. 2011.基于国家气象观测站逐日降水格点数据的交叉检验误差分析[J].高原气象, 30(6):1615-1625. http://www.cnki.com.cn/Article/CJFDTOTAL-GYQX201106020.htm

    Xiong Qiufen, Huang Mei, Xiong Minquan, et al. 2011. Cross-validation error analysis of daily gridded precipitation based on China meteorological observation[J]. Plateau Meteorology (in Chinese), 30(6):1615-1625. http://www.cnki.com.cn/Article/CJFDTOTAL-GYQX201106020.htm
    [21] 宇婧婧, 沈艳, 潘旸, 等. 2015.中国区域逐日融合降水数据集与国际降水产品的对比评估[J].气象学报, 73(2):394-410. doi: 10.11676/qxxb2015.033

    Yu Jingjing, Shen Yan, Pan Yang, et al. 2015. Comparative assessment between the daily merged precipitation dataset over China and the world's popular counterparts[J]. Acta Meterorologica Sinica (in Chinese), 73(2):394-410, doi: 10.11676/qxxb2015.033.
  • 加载中
图(15) / 表(2)
计量
  • 文章访问数:  1269
  • HTML全文浏览量:  7
  • PDF下载量:  2892
  • 被引次数: 0
出版历程
  • 收稿日期:  2016-07-20
  • 网络出版日期:  2016-11-09
  • 刊出日期:  2017-05-20

目录

    /

    返回文章
    返回