Potential fields methods, based on the exploitation of gravity and magnetic fields, are among the most important methods to recover fundamental information on the Earth crust structure at global, regional and local scales. The bottleneck for this kind of geophysical methods is often represented by the development of ad-hoc techniques to fully exploit the available data. In fact, each different technique can observe the effect of a single property of the subsurface and when we want to estimate this property from the observed field (the so-called inverse problem), several problems such as non-uniqueness and instability arise. A possible solution to these problems consists in jointly inverting, in a consistent way, different observed fields, possibly also incorporating all the available geological constraints. In the current work, we present an innovative Bayesian algorithm aimed at performing a full 3D joint inversion of gravity and magnetic fields constrained by geological a-priori qualitative information. The algorithm is tested on a real-case scenario, namely, a local study to estimate a complete 3D model of the Oka carbonatite complex. This complex is a composite pluton in Quebec (Canada), important for mining operations related to critical raw material such as Niobium and other rare earth. This example shows the reliability of the developed inversion algorithm and gives hints on the fundamental role that potential fields can play in mining activities.
