Layer Constrained Composite Modelling and Inversion of Gravity Data

The inversion process for gravity interpretation has to treat the intrinsic ambiguities of the gravimetric data; for this reason, the possible investigated models need to be constrained in order to recover the uniqueness of the inversion process.

In this work, a novel approach to gravimetric inversion is presented. Geometric information on the existing structures, derived from other geophysical studies, are introduced in the form of interfaces and geobodies. In this way, independent regions are defines and discontinuities of the model are introduced. Furthermore, due to the integral nature of gravimetric method, it was possible to develop a composite forward modelling. A flexible geological-driven forward and inversion machine has been therefore created, allowing the automatic introduction of the geometric information and the a priori calibration, which is based on specific geological interpretation. The inversion is driven by logarithmic barriers, in order to keep inversion parameter inside physical range.

Results show that geometric information put very strong constraints to density inversion. Furthermore, the composite forward modelling makes the algorithm more powerful with respect to standard approaches. The logarithmic barriers play a fundamental role in solving the intrinsic gravimetric ambiguity inside a geophysical domain.