Two-Dimensional Inversion of Papua New Guinea Magnetotelluric Dataset Assuming Static Shift as a Gaussian Distribution

The Papua New Guinea Magnetotelluric dataset was analyzed by applying Groom-Bailey tensor decomposition, and a consistent strike direction of N66° W was determined. The dataset was approximated by two-dimensional impedances, and frequency independent twist and shear. The static shift parameters (local anisotropy and site gain) were determined using a two-dimensional inversion where static shifts were also part of the model parameters. The model misfit was simultaneously minimized together with the following two norms: (1) roughness norm of the model, and (2) static shift L2 norm. The trade-off parameters between the model misfit and these norms were determined so as to minimize the Akaike's Bayesian Information Criterion (ABIC).