Journal
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT
Volume 17, Issue 4, Pages 260-271Publisher
SPRINGER
DOI: 10.1007/s00477-002-0117-1
Keywords
data integration; multivariate geostatistics; vector random fields; conditional covariances; spectral simulation; multiple attributes
Ask authors/readers for more resources
Conditional component random fields (CC) based on Cholesky decomposition of the multivariate spectra are introduced in this study to develop a new method for conditional simulation of vector attributes in environmental and geological phenomena. The CC are independent random fields with covariance models obtained from projections and conditioning in the frequency domain. The approach is to simulate one attribute in the physical space and use the results to estimate the other attributes in the frequency domain. Then, a CC for the next attribute is simulated and projected on the other attributes. In general, any attribute is built as the sum of inverse Fourier transform of the orthogonal projection of previous simulated CC plus a last CC simulated in the physical space. This simulation approach continues in this fashion for several attributes and the order of them may be changed for different realizations. This method allows for data conditioning and simulation. A simplified version for intrinsically correlated random fields allows for an approach that avoids the frequency domain.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available