The Pawsey Supercomputing Centre


Bayesian Lithological Inversion

Airborne electromagnetic (AEM) data are used in minerals exploration and to aid construction of groundwater models. Typically, these data are inverted using deterministic methods, for parameters of interest which must be interpreted for geological structure.  In contrast, our novel approach employs a stochastic methodology which attempts to invert directly for geologically-meaningful parameters and their uncertainties. Such stochastic methodologies are required in order to quantify risk in exploration, and structural uncertainty in groundwater models. In both applications, quantitative interpretation can be used to optimise exploration programs. Our novel approach is a pragmatic compromise between exhaustive MCMC stochastic and point-sampling deterministic methods, and has the potential to obtain results in real time.

Comments are closed.