Two-dimensional inversion modeling of magnetotelluric (MT) synthetic data of a graben structure using SimPEG

Abstract

The magnetotelluric (MT) method is a passive exploration method in geophysics that utilizes natural electromagnetic waves as a signal source. MT operates in the frequency range of 10-5-106 hertz, designed to study the structure of the conductivity below the earth's surface with a depth range from several tens of meters to the upper mantle. In this paper, 2-dimensional inversion modeling is performed on MT synthetic data using the SimPEG software. First, forward modeling is done by making a 2-dimensional conductivity model in the form of a valley (graben), which aims to produce MT synthetic data in TE (transverse electric) mode, TM (transverse magnetic) mode, and a combination of TE mode and TM mode. Next, an inversion modeling is performed on the MT synthetic data by adding a 5% Gaussian noise and a 10-5 floor as data uncertainty to obtain a 2-dimensional conductivity inversion model. The final result can be validated by comparing the true model and the inversion model and between observational data (synthetic) and predictive predicted data. The results of this study provide a significant fit of the model and suitability of the data. The inversion quality is validated with an RMS Error for TE mode of 0.349%, TM mode of 0.348%, and a combination of TE and TM mode of 0.249%.