Resistivity inversion software comparison

The application of geophysics may assist in solving environmental, geotechnical and exploration problems. One commonly used method in such applications is electrical resistivity tomography (ERT). In order to interpret the results, inversion is needed to create models estimating distribution of resistivity in the ground. Several inversion programs are available for geophysicists to use. It is therefore important to investigate if different software's give similar results for a given data set. In this paper, three different inversion software's have been compared through inversion of the forward response of the same synthetic model. The programs used are Res2Dinv, Aarhusinv and BERT. Two synthetic forward models have been used. As expected, Res2Dinv generally tends to exaggerate structures vertically and Aarhusinv to exaggerate them vertically. BERT seems to be somewhere in between, but does not seem to resolve sharp features in the synthetic forward model, resulting in rounded anomalies. A square-like feature close to the ground surface in the forward model was well resolved by all inversion software. However, it was not possible to resolve inclined blocks in any of the inversions. Possible explanations might be poor data coverage at deeper depths in the model and model discretization that cannot resolve inclined geometries.