Covariance Tapering for Interpolation of Large Spatial Datasets

Interpolation of a spatially correlated random process is used in many scientific areas. The best unbiased linear predictor, often called a kriging predictor in geostatistical science, requires the solution of a (possibly large) linear system based on the covariance matrix of the observations. In this article, we show that tapering the correct covariance matrix with an appropriate compactly supported positive definite function reduces the computational burden significantly and still leads to an asymptotically optimal mean squared error. The effect of tapering is to create a sparse approximate linear system that can then be solved using sparse matrix algorithms. Monte Carlo simulations support the theoretical results. An application to a large climatological precipitation dataset is presented as a concrete and practical illustration.

[1]  J. Pasciak,et al.  Computer solution of large sparse positive definite systems , 1982 .

[2]  A. Balakrishnan,et al.  Spectral theory of random fields , 1983 .

[3]  Phil Diamond,et al.  Robustness of variograms and conditioning of kriging matrices , 1984 .

[4]  Loren D. Pitt.  SPECTRAL THEORY OF RANDOM FIELDS (Translation Series in Mathematics and Engineering) , 1984 .

[5]  S. Yakowitz,et al.  A comparison of kriging with nonparametric regression methods , 1985 .

[6]  J. J. Warnes,et al.  A sensitivity analysis for universal kriging , 1986 .

[7]  A. Yaglom.  Correlation Theory of Stationary and Related Random Functions I: Basic Results , 1987 .

[8]  Michael L. Stein,et al.  Asymptotically Efficient Prediction of a Random Field with a Misspecified Covariance Function , 1988 .

[9]  Mark S. Handcock,et al.  Some asymptotic properties of kriging when the covariance function is misspecified , 1989 .

[10]  Michael L. Stein,et al.  Bounds on the Efficiency of Linear Predictions Using an Incorrect Covariance Function , 1990 .

[11]  Michael L. Stein,et al.  Uniform Asymptotic Optimality of Linear Predictions of a Random Field Using an Incorrect Second-Order Structure , 1990 .

[12]  N. Cressie.  The origins of kriging , 1990 .

[13]  Charles R. Johnson,et al.  Topics in Matrix Analysis , 1991 .

[14]  Barry W. Peyton,et al.  Block sparse Cholesky algorithms on advanced uniprocessor computers , 1991 .

[15]  John R. Gilbert,et al.  Sparse Matrices in MATLAB: Design and Implementation , 1992, SIAM J. Matrix Anal. Appl..

[16]  Robert Haining,et al.  Statistics for spatial data: by Noel Cressie, 1991, John Wiley & Sons, New York, 900 p., ISBN 0-471-84336-9, US $89.95 , 1993 .

[17]  Michael L. Stein,et al.  A simple condition for asymptotic optimality of linear predictions of random fields , 1993 .

[18]  Zongmin Wu,et al.  Compactly supported positive definite radial functions , 1995 .

[19]  Holger Wendland,et al.  Piecewise polynomial, positive definite and compactly supported radial functions of minimal degree , 1995, Adv. Comput. Math..

[20]  Ross Ihaka,et al.  Gentleman R: R: A language for data analysis and graphics , 1996 .

[21]  Michael L. Stein,et al.  Efficiency of linear predictors for periodic processes using an incorrect covariance function , 1997 .

[22]  Holger Wendland,et al.  Error Estimates for Interpolation by Compactly Supported Radial Basis Functions of Minimal Degree , 1998 .

[23]  M. Stein.  Predicting random fields with increasing dense observations , 1999 .

[24]  T. Gneiting.  Correlation functions for atmospheric data analysis , 1999 .

[25]  Tilmann Gneiting,et al.  Radial Positive Definite Functions Generated by Euclid's Hat , 1999 .

[26]  S. Cohn,et al.  Ooce Note Series on Global Modeling and Data Assimilation Construction of Correlation Functions in Two and Three Dimensions and Convolution Covariance Functions , 2022 .

[27]  J. Chilès,et al.  Geostatistics: Modeling Spatial Uncertainty , 1999 .

[28]  P. Houtekamer,et al.  A Sequential Ensemble Kalman Filter for Atmospheric Data Assimilation , 2001 .

[29]  J. Whitaker,et al.  Distance-dependent filtering of background error covariance estimates in an ensemble Kalman filter , 2001 .

[30]  R. Beatson,et al.  Smooth fitting of geophysical data using continuous global surfaces , 2002 .

[31]  Stephen Billings,et al.  Interpolation of geophysical data using continuous global surfaces , 2002 .

[32]  Michael L. Stein,et al.  The screening effect in Kriging , 2002 .

[33]  T. Gneiting.  Compactly Supported Correlation Functions , 2002 .

[34]  Craig J. Johns,et al.  Infilling Sparse Records of Spatial Fields , 2003 .

[35]  Roger Koenker,et al.  SparseM: A Sparse Matrix Package for R , 2003 .

[36]  Alexander Gribov,et al.  Geostatistical Mapping with Continuous Moving Neighborhood , 2004 .

[37]  Tilmann Gneiting,et al.  Convolution roots of radial positive definite functions with compact support , 2004 .

[38]  M. Fuentes,et al.  Sensitivity of ecological models to their climate drivers: statistical ensembles for forcing. , 2006, Ecological applications : a publication of the Ecological Society of America.

[39]  W. R. Madych,et al.  An estimate for multivariate interpolation II , 2006, J. Approx. Theory.