MT regularized inversion based on the minimum support gradient stabilizing functional

Abstract

Magnetotelluric (MT) inversion is ill-posed. Normally we use regularization methods by adding stabilizing functional to get a stable solution. Previous study of traditional smoothness constraint model could not properly invert sharp electrical boundary, while the inversion results of stabilizers suitable for sharp boundary are heavily depend on focusing parameters. In this paper, we construct a new minimum support gradient stabilizing functional which combines smooth and sharp boundary inversion together, and apply it in the unified form of regularized inversion objective functional. Through the comparisons of 1-D, 2-D inversion results between this new stabilizer and other common stabilizers, show that this new minimum support gradient stabilizer could better delineate the sharp electrical boundary than traditional stabilizers while ensuring a stable inversion result in MT regularized inversion.