Influence of different array datasets on reliability of electrical resistivity tomography

Abstract

The goal of this work is to study how the reliability of inverse model depends on a few basic parameters, as the combination of potential spacing and dipolar distance and, consequently, the number of measurements and of current dipoles, considering also how error affects inversion. The number of current dipole used is crucial, when using multichannel resistivity-meters, because it determines the overall acquisition time. A systematic comparison is presented between four 2D resistivity models and their images, obtained by the inversion of synthetic datasets relating to four different arrays: dipole-dipole (DD), pole-dipole (PD), Wenner-Schlumberger (WS) and multiple gradient (MG). For DD, PD and WS arrays a progression of eight different datasets are considered, by increasing the number of current dipoles but obtaining approximately the same amount of measures, and so increasing the investigation time. For MG array a progression of six datasets is obtained by increasing the current dipoles and so the lateral coverage. The goal is to study how this affect the resolution and the reliability of the tomographic inversion, particularly in presence of buried structures. Both noise-free and noisy data have been calculated and inverted. The results are compared using quality parameters of the reliability of the inversion. These are calculated for each cell of the first inverse model, and subsequently a mean value is obtained for the entire section or for areas coinciding with the abnormal structures. The sequences are also tested with field data to assess the validity of the theoretical results.