Inverse Problem Resolution Using Geologically Conditioned Gradient-Based Regularization – Application to the Yerrida Bas
J. Giraud, M. Lindsay, O. Vitaliy, R. Martin, E. Pakyuz-Charrier and M. Jessell
Event name: 80th EAGE Conference and Exhibition 2018
Session: Potential Field Methods I
Publication date: 11 June 2018
Info: Extended abstract, PDF ( 762.44Kb )
Price: € 20
The qualitative integration of geological and geophysical data is a useful approach to constrain inversion that aims to improve interpretation and to reduce the geophysical data misfit. Here, we introduce an inversion approach that incorporates probabilistic geological information into geophysical inversion in the absence of the data necessary to derive petrophysical constraints. We develop the local conditioning of a minimum-gradient regularization function that utilizes statistical information from a Probabilistic Geological Model (PGM) which is derived from Monte-Carlo simulations perturbing field geological measurements. The use of the PGM allows us to support model updates in geologically uncertain areas while promoting consistent changes in well-constrained portions of the model. The methodology is applied using dip, strike and orientation measurements and gravity data acquired in the Yerrida Basin. The results obtained are characterised by (i) higher density contrast and inverted models that are easier to interpret and, (ii) an improved geophysical data fit. We conclude that by allowing inversion to update the model more freely in geologically uncertain areas, our methodology is capable of focusing inversion and to reduce interpretation uncertainty.