An efficient inversion for two-dimensional direct current resistivity surveys based on the hybrid finite difference–finite element method

For efficient inversion code, the forward modeling routine, the sensitivity calculation, and the inversion algorithm must be efficient. Here, the hybrid finite difference–finite element algorithm, which is fast and accurate even when the slope of the topography is greater than 45°, is used as the forward modeling routine to calculate the responses. The sensitivity calculation is adapted from the most efficient adjoint Green’s function technique. Both of these algorithms are then driven with the data space Occam’s inversion. This combination of modules makes it possible to obtain an efficient inversion code based on MATLAB for two-dimensional direct current (DC) resistivity data. To demonstrate its efficiency, numerical experiments with our code and with commercial software are performed on synthetic data and real field data collected in the western part of Thailand where limestone and cavities dominate the region. In general, our code takes substantially longer than the commercial code to run but converges to a solution with a lower misfit. The result shows that the efficiency of our code makes it practical for real field surveys.

► Our developed 2-D DC resistivity inversion code can handle extreme topography.
► It is driven by the efficient hybrid FD–FE forward modeling method.
► It converges to a solution with a lower misfit than the commercial code but slower.
► It is practical for real field surveys for both single and mixed arrays inversions.