Sensitivity of constrained joint inversions to geological and petrophysical input data uncertainties with posterior geological analysis
The integration of petrophysical data and probabilistic geological modelling in geophysical joint inversion is a powerful tool to solve exploration challenges. Models obtained from geologically and/or petrophysically constrained inversions are the result of complex interactions between correspondingly diverse data sets. Therefore, it is important to understand how non-geophysical input uncertainty impacts inverted models. In this paper, we propose to study the influence of uncertainty in geological and petrophysical measurements used to derive prior information and constraints onto geophysical inversion. Starting from geological field data from the Mansfield area (Victoria, Australia), we simulate low, medium and high uncertainty levels in geological measurements and petrophysical data, combined into a series of nine realistic case scenarios. This allows us to investigate the impact and propagation of uncertainty from non-geophysical measurements into geophysical inversion. We calculate misfit indicators and reconstruct lithological models a posteriori to analyse inversion results. We complement the examination of inverted models with topological analysis of lithological models in order to quantify the geological resemblance between the recovered and reference models. Our work reveals that the influence of uncertainty in geological measurements over the recovered lithological models is significantly stronger than it is for petrophysical data. Our posterior analysis indicates that intermediate petrophysical uncertainty provides optimum results.