Considering missing information in 3D weights of evidence modelling, an example from the Drummond Basin, Queensland, Australia

This work discusses an application of GEOPAT3D, a software tool designed to integrate indicator patterns in three-dimensional space by using the Bayes' rule of combination. Bayesian logic is used to combine spatial evidence as in the Weights of Evidence method, a geostatistical approach that has seen extensive application in GIS analysis in several fields of science, including medicine and geology. A case study is considered here to illustrate the software usage and discuss advantages and limitations encountered when performing this type of spatial analysis in 3D. A posterior probability model was developed by 3D Weights of Evidence analysis of a 3D data set representative of the northern part of the Drummond Basin. Scope of this analysis was to estimate favourability indexes indicating the likelihood of finding a mineral deposit of low-sulphidation style in the selected region. Recent results led to the inclusion of a MI (Missing Information) class, which improves OT and NOT test results for the statistical acceptance of the CI (Conditional Independence) level of violation.