<![CDATA[Magnetotellurics (MT) is a passive geophysical method that relies on the principles of electromagnetic induction to delineate subsurface resistivity distributions. This study implements the Neighborhood Algorithm (NA) for one-dimensional (1D) inversion of MT data to overcome local minima traps and analysis model uncertainty. The performance of the NA is evaluated using the Rosenbrock function and synthetic MT data. The evaluation results show that the NA successfully achieves convergence with a misfit of less than 2.0 on synthetic data with 5% added Gaussian noise. Furthermore, the NA is applied to field MT data from Kilauea Volcano, Hawaii. The performance of the NA in modeling the field data is compared to a conventional approach, namely the Levenberg-Marquardt (LM) algorithm. The field data modeling results demonstrate that the NA (misfit 1.99) outperforms the LM algorithm (misfit 2.52). Additionally, the NA provides supplementary information in the form of a posterior probability density that maps the model's uncertainty bounds. The resistivity model, supported by uncertainty analysis, strengthens the information at Kilauea Volcano regarding the presence of a low-resistivity (conductive) zone at a depth of less than 3 km, which is associated with a shallow magma chamber (partial melt) or saline fluids.]]>
Copyrights © 2026
