Inversão de dados magnetotelúricos 1D utilizando o algoritmo de otimização de Levenberg-Marquardt

Abstract

The magnetotelluric method is widely used in geophysics for mapping geological structures. Magnetotelluric data inversion is a complex process that involves recovering the electrical properties of the subsurface from surface-observed magnetotelluric data, which are obtained from measurements of natural variations in the Earth’s electromagnetic fields. In this study, the magnetotelluric data inversion technique is employed to estimate the parameters of the geophysical model that is the source of such data. This process is carried out using the Levenberg-Marquardt algorithm to refine and adjust the model by minimizing the misfit between the MT data observed in the field and the data calculated by the forward modeling technique. The 1D approach assumes that the electrical properties of the soil vary only with depth. Experiments were conducted using synthetic data with Gaussian noise and real data from the COPROD magnetotelluric database. The results obtained indicate that the Levenberg-Marquardt method can be a viable option for 1D magnetotelluric data inversion, performing better when applied to data captured at lower frequencies. It’s worth noting that, being a local optimization algorithm, it also exhibits low computational cost.