Transdimensional joint inversion of surface wave, refraction, and resistivity data using Modified Barnacles Mating Optimizer for near-surface investigations

Abstract

This study presents a joint inversion of active Rayleigh wave dispersion curves, refraction travel times, and Vertical Electrical Sounding (VES) data sets using a recently modified global optimizer, yielding reliable parameter estimates. The Modified Barnacles Mating Optimizer (MBMO) was built on strengthening the BMO algorithm by integrating several enhancements. Before the inversion studies, modal analyses regarding the objective function defined and sensitivity studies of model parameters were performed using a synthetic near-surface model. These analyses indicate that the inverse problem is highly complex and prone to erroneous solutions with significant uncertainties. Additionally, the model parameters exhibit varying sensitivity levels. Therefore, this problem can be solved with an algorithm that obtains a good trade-off between global exploration and local exploitation. The efficiency of MBMO was tested on synthetic anomalies and on four real-world data sets from T & uuml;rkiye. Additionally, the performances of MBMO were unbiasedly compared with the BMO and the standard Particle Swarm Optimization (PSO) algorithm. A multiple model space strategy (MMSS) was innovatively constructed to achieve the transdimensional inversion without significantly increasing the difficulty of navigating the model space. We discarded the traditional approach of inverting partial derivatives model parameters. The solutions obtained from the real-world data cases were interpreted with the findings of previous geophysical and geological data. Post-inversion uncertainty analyses were performed accordingly to evaluate the reliability of the solutions. The findings show that MBMO outperforms BMO and PSO in solving the given problem. In addition, joint inversion of seismic and resistivity data sets can recover subsurface information with greater reliability due to the incorporated complementary information. The effectiveness of MBMO is not significantly affected by the form of the defined objective function.