Abstract
Consider the problem of generating a realization y1 of a Gaussian random field on a dense grid of points Ω1 conditioned on field observations y2 collected on a sparse grid of points Ω2. An approach to this is to generate first an unconditional realization y over the grid Ω=Ω1 ∪ Ω2, and then to produce y1 by conditioning y on the data y2. As standard methods for generating y, such as the turning bands, spectral or Cholesky approaches can have various limitations, it has been proposed by M. W. Davis to generate realizations from a matrix polynomial approximations to the square root of the covariance matrix. In this paper we describe how to generate a direct approximation to the conditional realization y1, on Ω1 using a variant of Davis' approach based on approximation by Chebyshev polynomials. The resulting algorithm is simple to implement, numerically stable, and bounds on the approximation error are readily available. Furthermore we show that the conditional realization y1 can be generated directly with a lower order polynomial than the unconditional realization y, and that further reductions can be achieved by exploiting a nugget effect if one is present. A pseudocode version of the algorithm is provided that can be implemented using the fast Fourier transform if the field is stationary and the grid Ω1 is rectangular. Finally, numerical illustrations are given of the algorithm's performance in generating various 2-D realizations of conditional processes on large sampling grids.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Abramowitz, M., and Stegun, I. A., 1970, Handbook of mathematical functions: Dover, New York, 1046 p.
Alabert, F., 1987, The practice of fast conditional simulations through the LU decomposition of the covariance matrix: Math. Geology, v. 19, no. 5, p. 369–386.
Black, T. C., and Freyberg, D. L., 1990, Simulation of one-dimensional correlated fields using a matrix-factorization moving average approach: Math. Geology, v. 22, no. 1, p. 39–62.
Borgman, L., Taheri, M., and Hagan, R., 1984, Three-dimensional frequency-domain simulations of geological variables,in Verly, G., David, M., Journel, A. G., and Marechal, A., eds., Geostatistics for natural resource characterization. Part 2: Reidel, Dordrecht, The Netherlands, p. 517–541.
Chan, R. H., and Jin, X. Q., 1992, A family of block preconditionrs for block systems: SIAM Jour. Sci. Stat. Comput., v. 13, no. 5, p. 1218–1235.
Davis, M. W., 1987a, Generating large stochastic simulations—the matrix polynomial approximation method: Math. Geology, v. 19, no. 2, p. 99–107.
Davis, M. W., 1987b, Production of conditional simulations via the LU triangular decomposition of the covariance matrix: Math. Geology, v. 19, no. 2, p. 91–98.
Dietrich, C. R., 1993, Computationally efficient generation of Gaussian conditional simulations over regular sample grids: Math. Geology, v. 25, no. 4, p. 439–451.
Dietrich, C. R., 1994, A simple and efficient implementation of the turning bands method with exact line realizations: Water Resour. Res., in press.
Dietrich, C. R., and Newsam, G. N., 1993, A fast and exact method for multidimensional Gaussian stochastic simulations: Water Resour. Res., v. 29, no. 8, p. 2861–2869.
Dietrich. C. R., and Newsam, G. N., 1994a, Fast and exact method for multidimensional Gaussian stochastic simulations: extension to conditional simulations: Water Resour. Res., in press.
Dietrich, C. R., and Newsam, G. N., 1994b, Generating correlated Gaussian random fields by orthogonal polynomial approximations to the square root of the covariance matrix: Jour. Statist. Comput. Simul., v. 50, p. 91–109.
Fox, L., and Parker, I. B., 1968, Chebyshev polynomials in numerical analysis: Oxford Univ. Press, Oxford, 205 p.
Golub, G. A., and van Loan, C. F., 1989, Matrix computations (2nd ed.): Johns Hopkins Univ. Press, Baltimore, 729 p.
Gradshteyn, I. S., and Ryzhik, I. M. 1980, Table of integrals, series, and products: Academic Press. New York, 1086 p.
IMSL, 1991, International Mathematics and Statistics Libraries, Version 2.0, Houston, Texas, 1371 P.
Journel, A. G., and Huijbregts, C. T., 1990, Mining geostatistics: Academic Press, New York, 461 p.
Krzanowski, W. J., 1990, Principles of multivariate analysis: a user's perspective: Oxford Statistical Science Series 3, Oxford, 563 p.
Newsam, G. N., 1988, On the asymptotic distribution of the eigenvalues of discretizations of a compact operator,in Proc. Center for Mathematical Analysis, Vol. 17: Special Program on Inverse Problems, Australian National Univ., Canberra, p. 92–105.
Press, W. H., Flannery, B. P., Teukolsky, S. A., and Vetterling, W. T., 1986, Numerical recipes: the art of scientific computing: Cambridge Univ. Press, Cambridge, 179 p.
Stahl, H., 1993, Best uniform rational approximation of m to r on [0, 1]: Bull. Am. Math. Soc., v. 28, no. 1, p. 116–122.
Stewart, G. W., 1973, Introduction to matrix computations: Academic Press, New York, 441 p.
Tompson, A. F. B., Ababou, R., and Gelhar, L. W. 1989, Implementation of the three-dimensional turning bands field generator: Water Resour. Res., v. 25, no. 10, p. 2227–2243.
Wolfram, S., 1991, Mathematica: a system for doing mathematics by computer (2nd ed.): Addison-Wesley, Redwood City, California, 961 p.
Zimmerman, D. L., 1989, Computationally exploitable structure of covariance matrices and generalized covariance matrices in spatial models: Jour. Statist. Comput. Simul., v. 32, p. 1–15.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Dietrich, C.R., Newsam, G.N. Efficient generation of conditional simulations by chebyshev matrix polynomial approximations to the symmetric square root of the covariance matrix. Math Geol 27, 207–228 (1995). https://doi.org/10.1007/BF02083211
Received:
Accepted:
Issue Date:
DOI: https://doi.org/10.1007/BF02083211