A Bayesian approach to the interpretation of airborne electromagnetic surveys: Quantifying data errors, model assessment, and lithology classification

Analysis of uncertainty is a crucial, yet often overlooked, aspect of any geophysical inverse problem. For airborne electromagnetic (AEM) data, uncertainty analysis is compounded by the large volume of data that are typically acquired in a survey. Here, we describe a Bayesian Markov chain Monte Carlo (McMC) algorithm initially developed for the analysis of frequency-domain electromagnetic data, including AEM, along with examples where this algorithm has been used to add new insight into model uncertainty. Recent algorithm developments will also be presented, including capabilities to assess random or systematic data errors as unknown parameters, simultaneously run multiple soundings in parallel allowing for the analysis of large surveys, speed up and assess convergence of the McMC algorithm, and to implement these methods for either time- or frequency-domain datasets.