Date of Award

2002

Degree Type

Thesis

Degree Name

Bachelor of Science Honours

Faculty

Faculty of Communications, Health and Science

First Advisor

Dr Ute Mueller

Second Advisor

Lyn Bloom

Abstract

Conditional sequential simulation algorithms have been used in geostatistics for many years but we currently find new developments are being made in this field. This thesis presents two new direct sequential simulation with histogram reproduction algorithms and compares them with the efficient and widely used sequential Gaussian simulation algorithm and the original direct sequential simulation algorithm. We explore the possibility of reproducing both the semivariogram and the histogram without the need for a transformation to normal space, through optimising an objective function and placing linear constraints on the local conditional distributions. Programs from the GSLIB Fortran library are expanded to provide a simulation environment. An isotropic and an anisotropic data set are analysed. Both sets are positively skewed and the exhaustive data is available to define global target distributions and for comparing the cumulative distribution functions of the simulated values.

Share

 
COinS