Zen and the Art of Bayesian Geology

Mining companies put substantial resources and effort in to discovering, characterising and exploiting mineral deposits. At the heart of these efforts is the need to create and maintain a geological and geophysical model of an ore body based on a variety of sparse two and three dimensional measurements. Typically, these measurements are diverse, sparse (often very sparse) and usually provide only ambiguous interpretations of the three dimensional ore body structure.

Professor Hugh Durrant-Whyte describes a number of efforts to apply Bayesian methods to the problem of minerals discovery and characterisation. Bayesian methods explicitly estimate probability distributions on geological models and capture the ambiguity and uncertainty inherent in geological modelling. Bayesian methods also provide a natural way of combining very diverse measurement types with typical prior expert data. Together, these methods can offer a coherent and insightful approach to framing and addressing the mineral exploitation problem.

Presented by the ARC Centre of Excellence for Mathematical & Statistical Frontiers