A three-dimensional weights of evidence model for the Drummond basin in NE Queensland: quantitative assessment of controlling variables for epithermal Au

In this contribution we report results of a mineral potential mapping study of the northern part of the Drummond Basin and Anakie Inlier. These areas are well endowed with epithermal Au, containing economic mineral deposits such as Pajingo, Mount Coolon and Glen Eva, representing classical examples of low-sulphidation, adulariasericite type gold systems.\ud \ud WofE (Weights of Evidence) is a GIS approach that assesses the likelihood of finding a mineral deposit of the sought type, in the region of interest. The method uses fuzzy logic derived from expert systems to assign the Weights of Evidence necessary during the construction of binary predictors, which are subsequently combined using the Bayes rule of combination to produce a probability surface (G. Bonham-Carter, 1994, Geographic information systems for geoscientists: modelling with GIS, Pergamon, v. 13,398 p.). When constructing a regional scale, 3D mineral potential model substantial subsurface geological information is desirable. However, such subsurface data is often limited, rendering the extension of this methodology in 3D difficult.\ud \ud We show possible solutions to this problem by using a range of geophysical inversion tools to estimate depth of intrusions, but also using geochemical regression to depth based on a combination of Factor Analysis and 3D WofE. Experimental analysis on the geochemical component was based on selected tables queried form the Terra Search database (training dataset).\ud \ud The advantage of using a 3D WofE model can be debated. However, initial results show that the use of geophysical information improves predictors (e.g., magnetic WORMS help distinguishing between real edges and false edges of intrusions, often misleading because of the surface relationships with cover sediments). Preliminary results show that 3D WofE may also provide a way to reconstruct geochemical signals undercover