Simultaneous Optimization of Vertical Electrical Sounding and Magnetotelluric Data Using a Genetic Algorithm

Abstract

There are many research studies that highlight the benefit of combining more than one geophysical method to delineate the subsurface. However, only a small number of studies discuss the use of genetic algorithm to simultaneously invert magnetotelluric (MT) and vertical electrical sounding (VES) data. The purpose of this research is to evaluate the efficacy of using the genetic algorithm technique to simultaneously optimize and invert MT and VES data. For this study, a GA inversion code was written in MATLAB, consisting of two parts: a forward program and an inverse program. The inverse program has an inherent forward model, which it uses to produce an apparent resistivity (from the corresponding input model parameter). The goal of GA inversion is to generate the best model parameter (that is, thickness and resistivity) whose apparent resistivity curve matches the field data’s apparent resistivity curve or the synthetic data’s apparent resistivity curve. The forward and GA inversion programs were tested on synthetic and real MT and VES datasets. A total of 34 MT and 15 VES soundings were acquired from different geothermal fields in Tuscany, Italy. The theoretical apparent resistivity values of the model parameter are extremely similar to the measured (experimental) apparent resistivity values, according to analysis of the inversion data. This is indicated by the low root mean square error. Results from the simultaneous inversion of the synthetic MT and VES models revealed negligible (less than 0.4 percent) errors in resistivity and thickness for each layer. In both cases, the error recorded by the application on field VES data and field MT data was less than 5 percent and less than 19 percent, respectively. This shows that the GA inverse technique produces accurate estimations of subsurface characteristics.

Keywords:Simultaneous Inversion, Genetic Algorithm, Vertical Electrical Sounding, Magnetotelluric, Optimization method

DOI: 10.7176/JEES/13-7-05

Publication date:September 30^th 2023