Ridge-regression algorithm for gravity inversion of fault structures with variable density

We derive an analytical expression for gravity anomalies of an inclined fault with density contrast decreasing parabolically with depth. The effect of the regional background, particularly the interference from neighboring sources of a fault structure, is ascribed by a polynomial equation. We have developed an inversion technique employing the ridge-regression iterative algorithm to infer the shape parameters of the fault structure, in addition to the effect of regional background. We demonstrate the validity of the proposed technique by inverting a gravity anomaly of a theoretical model, both with and without adding a regional background. The technique is insensitive to the effect of regional background. Two density-depth models of the Godavari subbasin in India are used in our interpretation of the gravity anomalies of the Ahiri-Cherla master fault. The interpreted results of a parabolic density model are found to be more geologically reasonable in comparison with the constant density model. The variations of the misfit function of the theoretical and observed gravity anomalies, the damping factor, and the shape parameters of the fault against the iteration number indicate the reliability of the interpretation.