评估集合卡曼滤波反演土壤湿度廓线的性能

doi:10.11867/j.issn.1001-8166.2010.04.0400

地球科学进展 ›› 2010, Vol. 25 ›› Issue (4): 400 -407. doi: 10.11867/j.issn.1001-8166.2010.04.0400

评估集合卡曼滤波反演土壤湿度廓线的性能

苟浩锋 ^1,2,刘彦华 ¹,张述文 ^1*,李得勤 ¹

1.兰州大学大气科学学院，甘肃省干旱气候变化与减灾重点实验室, 甘肃兰州 730000;2.兰州市气象局，甘肃兰州 730020

收稿日期:2009-10-29 修回日期:2010-01-15 出版日期:2010-04-10
通讯作者: 张述文（1966-)，男，河南固始人，教授，主要从事陆面过程、数据同化算法研究. E-mail:zhangsw@lzu.edu.cn
基金资助:
国家自然科学基金项目“同化多源观测估算土壤湿度廓线”(编号：40775065);公益性行业（气象）科研专项“WRF-EnSRF四维同化业务预报系统关键技术研究”（编号：GYHY200806029）共同资助. 

Assessing the Performance of the Ensemble Kalman Filter for Soil Moisture Profile Retrieval

Gou Haofeng ^1,2,Liu Yanhua ¹, Zhang Shuwen ¹,Li Deqin ¹

1. Key Laboratory of Arid Climate Change and Reducing Disaster of Gansu Province, College of Atmospheric Sciences, Lanzhou University, Lanzhou 730000, China;2.Lanzhou Meteorological Bureau, Lanzhou 730020, China

Received:2009-10-29 Revised:2010-01-15 Online:2010-04-10 Published:2010-04-10
Contact: shuwen zhang E-mail:zhangsw@lzu.edu.cn
Supported by:
苟浩锋（1979），男，甘肃庆阳人，硕士研究生，主要从事陆面过程同化算法应用的研究.E-mail:gouhf07@lzu.cn

下载PDF ( 1301 )

集合卡曼滤波由于易于使用而被广泛地应用到陆面数据同化研究中，它是建立在模型为线性、误差为正态分布的假设上，而实际土壤湿度方程是高度非线性的，并且当土壤过干或过湿时会发生样本偏斜。为了全面评估它在同化表层土壤湿度观测来反演土壤湿度廓线的性能，特引入不需要上述假设的采样重要性重采样粒子滤波，比较非线性和偏斜性对同化算法的影响。结果显示：不管是小样本还是大样本，集合卡曼滤波都能快速、准确地逼近样本均值，而粒子滤波只有在大样本时才能缓慢地趋近|此外，集合卡曼滤波的粒子边缘概率密度及其偏度和峰度与粒子滤波完全不同，前者粒子虽不完全满足正态分布，但始终为单峰状态，而后者粒子随同化推进经历了单峰到双峰再到单峰的变化。 

关键词: 集合卡曼滤波, 粒子滤波, 土壤湿度, Richards方程

The ensemble Kalman filter is an easy to use, flexible, and efficient data assimilation algorithm widely used in Land Surface Data Assimilation System. It bases on the normality approximation of model error and observational error as well as the linearity assumption of the soil moisture errors between the near-surface observation and other deep layers. However, the soil moisture equation is highly nonlinear and also soil moisture can be highly skewed toward the wet or dry ends. To evaluate the effects of these approximations and the performance of the ensemble kalman filter (EnKF) in estimating soil moisture profile based on the near-surface soil moisture measurements, the results from the EnKF are compared with those obtained from a sequential importance resampling (SIR) particle filter that is one of nonlinear filters . The comparative results show: the EnKF can quickly and accurately obtain the exact soil moisture profile regardless of the small number of ensemble members used (40) or the large number of ensemble members used (800); however, the SIR needs the large number of replicates required to accurately represent the variable conditional probability densities. The near-surface soil moisture forecast densities, the skewness and kurtosis and obtained from the EnKF are completely different from those from the SIR filter; the densities from the EnKF is only one modal during the total assimilation time window but those from the SIR experiences from one mode to two modes and again to one mode process.

Key words: Ensemble Kalman filter, Sampling importance resampling, Soil moisture, Richards equation

中图分类号:

P461.4

[1] Walker J P, Willgoose G R, Kalma J D.One-dimensional soil moisture profile retrieval by assimilation of near-surface observations: A comparison of retrieval algorithms[J].Advances in Water Resources,2001,24(6):631-650.
[2] Entekhabi D,Nakamura H, Njoku E G. Solving the inverse problem for soil moisture and temperature profiles by sequential assimilation of multifrequency remotely sensed observations[J].IEEE Transactions on Geoscience and Remote Sensing,1994, 32:438-448.
[3] Reichle R H, Mclaughlin D B, Entekhabi D. Hydrologic data assimilation with the ensemble Kalman filter[J].Monthly Weather Review,2002,130:103-114.
[4] Hoeben R, Troch P A. Assimilation of active microwave observation data for soil moisture profile estimation[J].Water Resources Research,2000,36:2 805-2 819.
[5] Galantowicz J F, Entekhabi D, Njoku E G.Tests of sequential data assimilation for retrieving profile soil moisture and temperature from observed L-Band radiobrightness[J]. IEEE Transactions on Geoscience and Remote Sensing,1999,37:1 860-1 870.
[6] Xiwu Zhan, Houser P R, Walker J P, et al.A method for retrieving high-resolution surface soil moisture from hydros L-Band radiometer and radar observations[J].IEEE Transactions on Geoscience and Remote Sensing,2006,44:1 534-1 544.
[7] Boussetta S, Koike T, Graf T, et al. Development of a coupled land-atmosphere satellite data assimilation system for Improved local atmospheric simulations[J]. Remote Sensing of Environment,2008,112(3):720-734.
[8] Jones A S, Evic T V, Vonder Haar T H.A microwave satellite observational operator for variational data assimilation of soil moisture[J]. Journal of Hydrometeorology, 2004, 5:213-229.
[9] Tian Xiangjun ,Xie Zhenghui.A land surface soil moisture data assimilation framework in consideration of the model subgrid-scale heterogeneity and soil water thawing and freezing[J]. Science in China (Series D),2008, 51: 992-1 000.
[10] Han Xujun, Li Xin. Review of the nonlinear filters in the land data assimilation[J]. Advances in Earth Science, 2008,23(8):813-820.[韩旭军, 李新.非线性滤波方法与陆面数据同化[J].地球科学进展，2008,23(8):813-820.]
[11] Arulanmpalam S, Maskell S, Gordon N,et al. A tutorial on particle filters for on-line non-linear/non-gaussian bayesian tracking[J].IEEE Transactions on Signal Processing,2002,50:174-188.
[12] Zhou Yuhua, Mclaughlin D, Entekhabi D.Assessing the performance of the ensemble Kalman filter for land surface data assimilation[J]. Monthly Weather Review，2005,134:2 128-2 142.
[13] Zhang Qiang, Wang Sheng. On land surface processes and its experimental study in Chinese Loess Plateau[J]. Advances in Earth Science,2008，2:167-173.[张强，王胜.关于黄土高原陆面过程及其观测试验研究[J].地球科学进展，2008，2：167-173.]
[14] Zgang Qiang,Hu Xiangjun, Wang Sheng,et al. Some technological and scientific issues about the experimental study of land surface processes in Chinese Loess Plateau(LOPEX)[J]. Advances in Earth Science,2009，4:363-371.[张强,胡向军,王胜,等.黄土高原陆面过程试验研究(LOPEX)有关科学问题[J].地球科学进展，2009,4:363-371.].
[15] Xu Meng,Wang Hongwei,Hu Shiqiang. Optimal sensor decision based on particle filter[J].Journal of Shanghai Jiaotong University(Science), 2006,3:296-300.
[16] Peng Zhenming, Li Yalin,Wei Wen′ge,et al. Nonlinear AVO inversion using particle filter[J]. Chinese Journal of Geophysics, 2008,4:1 218-1 225.[彭真明，李亚林,魏文阁，等.粒子滤波非线性AVO反演方法[J].地球物理学报,2008,4:1 218-1 225]. 
[17] Wu Baocheng. Research and Application of Particle Filter Resampling Algorithms[D]. Harbin：Harbin University of Technology,2006.[吴宝成.粒子滤波重采样算法研究及其应用[D].哈尔滨：哈尔滨工业大学，2006.]
[18] Fan Dianhua. Particle filter[J]. Journal of the Graduates Sun Yat—Sen University, 2005,2:22-32.[范典华.粒子滤波[J].中山大学研究生学刊, 2005,2:22-32.]
[19] Snyder C, Bengtsson T, Bickel P. Obstacles to high-dimensional particle filtering[J]. Monthly Weather Review,2008,136:4 629-4 640.
[20] Evensen G. Sequential data assimilation with a nonlinear QG model using Monte Carlo methods to forecast error statistics[J]. Journal of Geophysical Research, 1994,99:10 143-10 162. 
[21] Burgers G, van Leeuwen P J, Evensen G. Analysis scheme in the ensemble Kalman filter[J]. Monthly Weather Review, 1998,126:1 719-1 724.
[22] Whitaker J S, Hamill T M. Ensemble data assimilation without perturbed observations[J]. Monthly Weather Review, 2002,130:1 913-1 924.
[23] Dai Y J, Zeng X B, Dickinson R E I, et al. The common land model (CLM)[J]. Bulletin of the American Meteorological Society,2003,84, 4: 1013-1023,doi:10.1175/BAMS-84-8-1013.
[24] Oleson K W, Coauthors. Technical description of the community land model (CLM)[C]//NCAR Technical Note NCAR/TN-461_STR, 2004:186.
[25] Zhao Aihui,Huang Mingbin,Shi Zhuye. Evaluation of parameter models for estimating loess soil particle-size distribution[J]. Transactions of the CSAE,2008,1:1-6.[赵爱辉，黄明斌，史竹叶. 土壤颗粒分布参数模型对黄土性土壤的适应性研究[J].农业工程学报,2008,1:1-6.]
[26] Wang Xiahui, Wang Yiquan, Kuznetsov M S. Study on physical Properties of several main soils in Loess Plateau[J].Journal of Soil and Water Conservation,2000,4:99-103.[王夏晖,王益权, Kuznetsov M S. 黄土高原几种主要土壤的物理性质研究[J].水土保持学报,2000,4:99-103.]

[1]	李得勤, 张述文, 文小航, 贺慧. 土壤湿度参数化及对天气和气候模拟影响的研究进展[J]. 地球科学进展, 2016, 31(3): 236-247.
[2]	兰鑫宇, 郭子祺, 田野, 雷霞, 王婕. 土壤湿度遥感估算同化研究综述[J]. 地球科学进展, 2015, 30(6): 668-679.
[3]	毛伏平, 张述文, 叶丹, 杨茜茜. 模式时间关联误差对集合平方根滤波估算土壤湿度的影响[J]. 地球科学进展, 2015, 30(6): 700-708.
[4]	朱忠礼,林柳莺，徐同仁. 海河流域不同下垫面土壤水分动态模拟研究[J]. 地球科学进展, 2012, 27(7): 778-787.
[5]	李得勤，段云霞，张述文. 土壤湿度观测、模拟和估算研究[J]. 地球科学进展, 2012, 27(4): 424-434.
[6]	李新,摆玉龙. 顺序数据同化的Bayes滤波框架[J]. 地球科学进展, 2010, 25(5): 515-522.
[7]	王晓婷,郭维栋,钟中,崔晓燕. 中国东部土壤温度、湿度变化的长期趋势及其与气候背景的联系[J]. 地球科学进展, 2009, 24(2): 181-191.
[8]	韩旭军，李新. 非线性滤波方法与陆面数据同化[J]. 地球科学进展, 2008, 23(8): 813-820.
[9]	关止,赵凯,宋冬生. 利用反射GPS信号遥感土壤湿度[J]. 地球科学进展, 2006, 21(7): 747-750.
[10]	张生雷,谢正辉,田向军,师春香,陈锋. 基于土壤水模型及站点资料的土壤湿度同化方法[J]. 地球科学进展, 2006, 21(12): 1350-1362.
[11]	马柱国，符淙斌，谢力，陈文海，陶树望. 土壤湿度和气候变化关系研究中的某些问题[J]. 地球科学进展, 2001, 16(4): 563-566.
[12]	马柱国,魏和林,符淙斌. 土壤湿度与气候变化关系的研究进展与展望[J]. 地球科学进展, 1999, 14(3): 299-305.

阅读次数

全文

摘要