hide
Free keywords:
Active Constraint Balancing (ACB), Inversion, Regularization parameter, Magnetotelluric (MT), Model parameter
Abstract:
Inversion of magnetotelluric (MT) data is a non-linear ill-posed inverse problem, and is commonly solved with an iterative linearized approach. In order to solve the minimization problem a regularization parameter is employed to balance between the norm of the data misfit and the norm of the model. Determination of a suitable regularization parameter is necessary to achieve both resolution and stability in inversion. In most inversion schemes, the regularization parameter is
applied globally to the entire model. In this study, we test the apability of active constraint balancing (ACB) approach, which was introduced by Yi et al. (2003). The approach determines a spatially varying regularization parameter via Backus-Gilbert spread function analysis for model parameters. Here, we use the 2D MT inversion code developed by Lee et al., (2009) to test the ACB on 2D synthetic and field MT datasets.
