Three-dimensional impedance tensor correction in the presence of out of quadrant phases and non-inductive distortions

The non-inductive effects of structures with dimensions smaller than the measurement scale and at shallow depths prevent the correct observation of the regional electrical conductivity model and therefore make the modeling and interpretation of magnetotelluric data difficult and in some cases unreliable. The solutions that have been presented to estimate the intensity of these distortions and recover regional information have been mostly focused on two-dimensional modeling. Several studies have shown the adverse effects of galvanic distortions on 3D magnetotelluric inversion results. Removing or correcting these distortions in practice is, however, rarely done before 3D inversion due to the extreme under-determination of the problem of recovering non-distorted or regional information in 3D environments need to apply more constraints. In this research, the complexity of magnetotelluric data in the Sablan geothermal area, in the northwest of Iran, was studied. By fitting the 3D/2D/3D model in the region, shear and twist parameters have been evaluated for a part of the period interval in which the data show 2D behavior, according to skew angle values. In the next step, the same distortion parameters were applied to the three-dimensional part of the data and the components of the impedance tensor for the 3D structure were recovered. For this purpose, the phase tensor (Caldwell et al, 2004), the rotational invariants of the magnetotelluric tensor (Weaver et al, 2000) and the approach presented by Ledo et al (1998) have been used. In order to correct the distortion, in addition to the estimated values for the twist and shear angles and the period interval selected for matching the two-dimensional model, it should also be taken into account that in the two-dimensional model, the values of the distortion parameters, i.e. the torsion and shear angles, remain constant with changing period. With this criterion, despite the values of the skew angle showing a two-dimensional behavior, the average distortion parameters for a number of stations could not be selected due to high fluctuations. It seems notable to emphasize that the skew parameter is the only necessary and not sufficient condition to confirm the two-dimensional situation. The magnitude and phase of all the components of the recovered impedance tensor are different from before, indicating the importance of the distortion correction procedure before the 3D modeling and inversion. In addition, in more than half of the examines magnetotelluric sites, impedance tensor phase components are not located in the corresponding trigonometric quadrants that are constrained to be placed in them in normal 2D or even 3D earth conditions. There are numerous examples of observing and studying these effects in magnetotelluric data. This behavior is attributed to factors such as anisotropy, three-dimensional complexities, two-dimensional structures with large resistivity contrast and severe distortion (Egbert, 1990; Pina and Dentith, 2018; Wannamaker, 2005; Jones et al, 1998). The results have shown that these abnormal values are related to the rotation angle and distortion level. Some of these stations show very large distortion angles.