A multi-layer soil moisture data assimilation using support vector machines and ensemble particle filter

Summary Hybrid data assimilation (DA) is greatly used in recent hydrology and water resources research. In this study, one newly introduced technique, the ensemble particle filter (EnPF), formed by coupling ensemble Kalman filter (EnKF) with particle filter (PF), is applied for a multi-layer soil moisture prediction in the Meilin watershed based on the support vector machines (SVMs). The data used in this paper includes six-layer soil moisture: 0–5 cm, 30 cm, 50 cm, 100 cm, 200 cm and 300 cm and five meteorological parameters: soil temperature at 5 cm and 20 cm, air temperature, relative humidity and solar radiation in the study area. In order to investigate this EnPF approach, another two filters, EnKF and PF are applied as another two data assimilation methods to conduct a comparison. In addition, the SVM model simulated data without updating with data assimilation technique is discussed as well to evaluate the data assimilation technique. Two experimental cases are explored here, one with 200 initial training ensemble members in the SVM training phase while the other with 1000 initial training ensemble members. Three main findings are obtained in this study: (1) the SVMs machine is a statistically sound and robust model for soil moisture prediction in both the surface and root zone layers, and the larger the initial training data ensemble, the more effective the operator derived; (2) data assimilation technique does improve the performance of SVM modeling; (3) EnPF outweighs the performance of other two filters as well as the SVM model; Moreover, the ability of EnPF and PF is not positively related to the resampling ensemble size, when the resampling size exceeds a certain amount, the performance of EnPF and PF would be degraded. Because the EnPF still performs well than EnKF, it can be used as a powerful data assimilation tool in the soil moisture prediction.

[1]  Jing Zhao,et al.  Particle filter based on Particle Swarm Optimization resampling for vision tracking , 2010, Expert Syst. Appl..

[2]  Xianhong Xie,et al.  Data assimilation for distributed hydrological catchment modeling via ensemble Kalman filter , 2010 .

[3]  Mac McKee,et al.  Multi-time scale stream flow predictions: The support vector machines approach , 2006 .

[4]  M. McKee,et al.  SOIL MOISTURE PREDICTION USING SUPPORT VECTOR MACHINES 1 , 2006 .

[5]  I-Fan Chang,et al.  Support vector regression for real-time flood stage forecasting , 2006 .

[6]  J. Famiglietti,et al.  Improving parameter estimation and water table depth simulation in a land surface model using GRACE water storage and estimated base flow data , 2010 .

[7]  M. Ferrante,et al.  Particle filtering approximations for a Gaussian-generalized inverse Gaussian model , 2009 .

[8]  Elaine Martin,et al.  Particle filters for state and parameter estimation in batch processes , 2005 .

[9]  Yuei-An Liou,et al.  Retrieving soil moisture from simulated brightness temperatures by a neural network , 2001, IEEE Trans. Geosci. Remote. Sens..

[10]  A. W. Heemink,et al.  The ensemble particle filter (EnPF) in rainfall-runoff models , 2009 .

[11]  Mac McKee,et al.  Support vectors–based groundwater head observation networks design , 2004 .

[12]  Jasmeet Judge,et al.  Effect of simultaneous state–parameter estimation and forcing uncertainties on root-zone soil moisture for dynamic vegetation using EnKF , 2010 .

[13]  Rajib Maity,et al.  Potential of support vector regression for prediction of monthly streamflow using endogenous property , 2010 .

[14]  W. Crow,et al.  The assimilation of remotely sensed soil brightness temperature imagery into a land surface model using Ensemble Kalman filtering: a case study based on ESTAR measurements during SGP97 , 2003 .

[15]  Shie-Yui Liong,et al.  FLOOD STAGE FORECASTING WITH SUPPORT VECTOR MACHINES 1 , 2002 .

[16]  Niko E. C. Verhoest,et al.  Improvement of TOPLATS‐based discharge predictions through assimilation of ERS‐based remotely sensed soil moisture values , 2002, Hydrological Processes.

[17]  Bernhard Schölkopf,et al.  Learning with kernels , 2001 .

[18]  Wade T. Crow,et al.  A land surface data assimilation framework using the land information system : Description and applications , 2008 .

[19]  A. Kalra,et al.  Using oceanic‐atmospheric oscillations for long lead time streamflow forecasting , 2007 .

[20]  Ted Engman,et al.  Why bother for 0.0001% of Earth's water? Challenges for soil moisture research , 1996 .

[21]  Vladimir Vapnik,et al.  Statistical learning theory , 1998 .

[22]  Neil J. Gordon,et al.  Editors: Sequential Monte Carlo Methods in Practice , 2001 .

[23]  Dennis K. McLaughlin,et al.  Opportunities for enhanced collaboration within the data assimilation community , 2005 .

[24]  Geir Evensen,et al.  The Ensemble Kalman Filter: theoretical formulation and practical implementation , 2003 .

[25]  Bernhard Schölkopf,et al.  Comparing support vector machines with Gaussian kernels to radial basis function classifiers , 1997, IEEE Trans. Signal Process..

[26]  Pao-Shan Yu,et al.  Statistical downscaling of daily precipitation using support vector machines and multivariate analysis , 2010 .

[27]  Gwo-Fong Lin,et al.  Effective forecasting of hourly typhoon rainfall using support vector machines , 2009 .

[28]  Peter Jan,et al.  Particle Filtering in Geophysical Systems , 2009 .

[29]  Craig J. Johns,et al.  A two-stage ensemble Kalman filter for smooth data assimilation , 2008, Environmental and Ecological Statistics.

[30]  Anthony W. England,et al.  A land surface process/radiobrightness model with coupled heat and moisture transport for prairie grassland , 1999, IEEE Trans. Geosci. Remote. Sens..

[31]  D. McLaughlin,et al.  Hydrologic Data Assimilation with the Ensemble Kalman Filter , 2002 .

[32]  Nello Cristianini,et al.  An Introduction to Support Vector Machines and Other Kernel-based Learning Methods , 2000 .

[33]  Luis A. Bastidas,et al.  Downscaling and Assimilation of Surface Soil Moisture Using Ground Truth Measurements , 2008, IEEE Transactions on Geoscience and Remote Sensing.

[34]  Zhongbo Yu,et al.  Data assimilation using support vector machines and ensemble Kalman filter for multi-layer soil moisture prediction , 2010 .

[35]  Wade T. Crow,et al.  Towards the estimation root-zone soil moisture via the simultaneous assimilation of thermal and microwave soil moisture retrievals , 2010 .

[36]  H. Moradkhani.  Hydrologic Remote Sensing and Land Surface Data Assimilation , 2008, Sensors.

[37]  Georges Oppenheim,et al.  The particle filters and their applications , 2008 .

[38]  Ronggao Liu,et al.  Simultaneous estimation of both soil moisture and model parameters using particle filtering method through the assimilation of microwave signal , 2009 .

[39]  Mac McKee,et al.  Soil Moisture Data Assimilation Using Support Vector Machines and Ensemble Kalman Filter 1 , 2007 .

[40]  Nando de Freitas,et al.  Sequential Monte Carlo Methods in Practice , 2001, Statistics for Engineering and Information Science.

[41]  Vladimir N. Vapnik,et al.  The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[42]  Michael Isard,et al.  CONDENSATION—Conditional Density Propagation for Visual Tracking , 1998, International Journal of Computer Vision.

[43]  G. Evensen.  Sequential data assimilation with a nonlinear quasi‐geostrophic model using Monte Carlo methods to forecast error statistics , 1994 .

[44]  A. Khalil,et al.  th International Conference on Information Fusion ( FUSION ) New applications for Information Fusion and Soil Moisture Forecasting , 2006 .

[45]  Osama Z. Al-Hamdan,et al.  Soil Moisture Profile Development from Surface Observations by Principle of Maximum Entropy , 2010 .

[46]  Dimitri P. Solomatine,et al.  Model Induction with Support Vector Machines: Introduction and Applications , 2001 .

[47]  A. Weerts,et al.  Particle filtering and ensemble Kalman filtering for state updating with hydrological conceptual rainfall‐runoff models , 2006 .

[48]  P. Leeuwen,et al.  Nonlinear data assimilation in geosciences: an extremely efficient particle filter , 2010 .

[49]  Joaquín Míguez,et al.  Analysis of parallelizable resampling algorithms for particle filtering , 2007, Signal Process..

[50]  Luis A. Bastidas,et al.  Multiobjective particle swarm optimization for parameter estimation in hydrology , 2006 .

[51]  Amilcare Porporato,et al.  A review of soil moisture dynamics: From rainfall infiltration to ecosystem response , 2005 .

[52]  Emad A. El-Sebakhy,et al.  Forecasting PVT properties of crude oil systems based on support vector machines modeling scheme , 2009 .

[53]  F. Girosi,et al.  Nonlinear prediction of chaotic time series using support vector machines , 1997, Neural Networks for Signal Processing VII. Proceedings of the 1997 IEEE Signal Processing Society Workshop.

[54]  J. Mather,et al.  Climatology: fundamentals and applications , 1974 .

[55]  D. Entekhabi,et al.  Land data assimilation and estimation of soil moisture using measurements from the Southern Great Plains 1997 Field Experiment , 2001 .