Date of Award


Degree Type


Degree Name

Bachelor of Science Honours


Faculty of Communications, Health and Science

First Advisor

Dr Lyn Bloom

Second Advisor

Dr Ute Mueller


In the earth sciences, and particularly in the mining of precious metals, data distributions are often strongly positively skewed. When making decisions on the potential profitability of a gold mine, for example, the high values of the distribution are of particular importance. Indicator kriging provides estimates of cumulative distribution functions from which grade tonnage curves may be calculated. Multiple or full indicator kriging requires a semivariogram to be modelled and a kriging system of equations to be solved for each cut off. This can be time consuming and modelling indicator semivariograms at high cut-offs may be difficult because of the low number of data above the cut off. One way to avoid these problems is to use median indicator kriging. Median indicator kriging uses the same semivariogram model at each cut off and may not perform as well as full indicator kriging. This thesis presents comparisons of median indicator kriging and full indicator kriging in the analysis of three suites of data for which the assumptions of median indicator kriging are only approximately satisfied. The distributions of the sample data sets have different degrees of skewness and sparseness. Two of the data suites represent highly skewed gold mineralisations and grade tonnage curves are obtained from the results of both indicator kriging methods. The third suite consists of approximately normally distributed air permeability data. The results from the two methods are compared with reality and show that median indicator kriging performs as well as full indicator kriging in each case.