Gravity inversion of basement relief using Particle Swarm Optimization by automated parameter selection of Fourier coefficients

Abstract

A global optimization technique using particle swarm optimization (PSO) is presented to estimate the depth to the basement of a heterogeneous sedimentary basin from the vertical component of the residual gravity anomalies. Inversion of basement relief for known density distributions has particular importance in many real applications such as mineral exploration, geothermal exploration, etc. Generally, gravity inversion in the Fourier domain allows a reduction in the optimization parameter. However, in the present study, an automated parameter selection criterion is developed for further reduction of optimizing parameters. A detailed uncertainty appraisal analysis is also performed for different configurations of synthetic models, which ensures the reliability of the optimization technique. All synthetic models are contaminated with white Gaussian noise, and an optimized depth profile is compared with noise-free data for sensitivity analysis. The result shows the robustness of the optimization method in the presence of noise. The technique is implemented on two real gravity anomaly profiles (1) Godavari basin, India, and (2) Sayula basin, Mexico. The optimized depth profile shows a good agreement with the published results using other optimization techniques. The method developed in the present work is a novel approach for the automatic selection of parameters as per the model’s complexity. It provides an inversion technique that considers a small number of parameters with a minimum computational expense.