Computation of spectral parameters from time‐domain induced polarization data for mineral identification

Abstract

Abstract Induced polarization (IP) effect is widely used in the search of disseminated minerals all over the world. Spectral parameters computed from time‐domain IP data play a significant role in characterizing the sources, but mineral identification still remains a challenge. In this paper, the Levenberg–Marquardt method of inversion is adopted in estimating the spectral parameters from time‐domain IP data to identify the polarizable sources. The algorithm is tested on synthetic time‐domain IP data for its robustness to variable noise levels. Model sensitivity analyses on synthetic data were also studied with respect to acquisition time and each individual model parameter. Error analyses on extracted parameters indicated that these are well resolved and correlated if the relaxation time is within the range of acquisition time. The parameters remain poorly resolved/unresolved for smaller values of chargeability and frequency dependence. The algorithm has also been tested over known case histories of time‐domain IP data and compared the estimated spectral parameters with those of published results. The inferences drawn from computed spectral parameters on field‐observed IP transients are in good correlation with other data sets and borehole information. The methodology has successfully shown its usefulness in identifying large polarizable sources occurring at shallow levels from time‐domain IP data.