DATA RETRIEVAL AND MULTIDIMENSIONAL SIMULATION OF MINERAL RESOURCES

Two subjects are studied in this thesis. First, a computer interactive data-retrieval system for mineral resources is developed. Secondly, a statistical simulation technique is presented which produces Gaussian simulation having zero mean and desired covariance function in two- and three-dimensional space.; The mineral resource data retrieval (MRDR) system works interactively in a simple question-and-answer routine. Mineral engineers and geologists with very little knowledge of computer use are able to interact with this system and retrieve data-information based on large numbers of user-selected criteria.; The statistical properties of fast Fourier transform coefficients for two-dimensional processes are studied in detail. For Gaussian processes the real and imaginary parts of the Fourier coefficients are found to be normally distributed and are statistically independent from frequency to frequency. In Chapter IV the spectral density functions corresponding to the multidimensional covariance functions with elliptical zone of influence and constant or varying nugget effect are extensively discussed. Also some common covariance functions are considered and their corresponding spectral density functions are derived.; Finally, two- and three-dimensional simulation of mineral resources using a direct frequency-domain method is presented. This method gives the opportunity to work in real two- and three-dimensional space directly and efficiently with very little computer cost. The method can be used also to simulate a realistic behavior of a deposit, since it has the potential to produce the important characteristics of reality observed; that is, autocorrelation (covariance of variogram) of the variables in a deposit.