Routine 2D inversion of magnetotelluric data using the determinant of the impedance tensor

Recent developments in the speed and quality of data acquisition using the radiomagnetotelluric (RMT) method, whereby large amounts of broadband RMT data can be collected along profiles, have prompted us to develop a strategy for routine inverse modeling using 2D models. We build a rather complicated numerical model containing both 2D and 3D elements believed to be representative for shallow conductors in crystalline basement overlain by a thin sedimentary cover. We then invert the corresponding synthetic data on selected profiles, using both traditional NIT approaches, as well as the proposed approach, which is based on the determinant of the MT impedance tensor. We compare the estimated resistivity models with the true models along the selected profiles and find that the traditional approaches often lead to strongly biased models and bad data fit, in contrast to those using the determinant. In this case, much of the bias is removed and the data fit is improved. The determinant of the impedance tensor is independent of the chosen strike direction, and once the a priori model is set, the best fitting model is found to be practically independent of. the starting model used. We conclude that the determinant of the impedance tensor is a useful tool for routine inverse modeling.