View article

There are different Geostatistical methods for generating realizations. Each of these methods has different advantages and disadvantage. As an example Cholesky decomposition is based on decomposing covariance matrix for generating realizations. When the dimension of problem is large, decomposing covariance matrix would be very difficult. This paper shows procedure for generating realizations using moving average method which is similar to the Cholesky decomposition method. In this method, decomposing numerical covariance matrix can be avoided by finding square root of covariance matrix analytically. This can be done using Fourier transform properties. Using this method, generating realizations from large filed could be possible in a reasonable time. This paper shows the procedure by considering different examples.