Modified symbiotic organisms search (SOS) algorithm for controlled-source audio-frequency magnetotellurics (CSAMT) one-dimensional (1D) modelling

Abstract: A relatively new optimisation algorithm called symbiotic organisms search (SOS), which mimics survival efforts of organisms in an ecosystem, is presented. The interactions or synergy among organisms for survival involving mutualism, commensalism and parasitism phases are simulated as the stochastic search mechanism for optimum models in the search space. The SOS algorithm has minimal tuning parameters compared to other nature-inspired optimisation algorithms, i.e., the number of population and the number of generations. We slightly modified the original algorithm to achieve a better balance between exploration and exploitation of the search space. We illustrate the proposed method with inversions of controlled-source audio-frequency magnetotellurics (CSAMT) data to obtain one-dimensional (1D) resistivity model representing the horizontally layered subsurface. Application to CSAMT synthetic data showed promising results in terms of synthetic model recovery and data fitting. To further validate the inversion method, we invert CSAMT field data from Lahendong geothermal field located in North Sulawesi, Indonesia. More specifically, inversion results of CSAMT data from representatives sounding stations well explain the geothermal model inferred from the nearby borehole data. Research highlights: Relatively new nature-inspired optimisation algorithm applied for inversion of geophysical data, namely CSAMT data inversion for 1D model.Modification of the original SOS algorithm to improve convergence.Application of the SOS algorithm to real-CSAMT data from a geothermal field.