Inversion of Induced Polarization Parameters From Time‐Domain Electromagnetic Data Using the Debye Decomposition Model

ABSTRACT

Time‐domain electromagnetics (TEM) is a valuable tool for exploring seafloor hydrothermal deposits, which are characterized by variations in resistivity and chargeability. Several surveys using the Waseda integrated seafloor time‐domain electromagnetics (WISTEM) have been conducted. Negative transients, which are attributed to induced polarization (IP) effects for this coincident‐type transmitter–receiver configuration loop, have been observed in data collected over a known deposit in the Okinawa Trough (2018). The Cole–Cole type model is a common parameterization of complex resistivity due to IP effects; however, it implicitly assumes that the data are sensitive to a wide range of frequencies. TEM data have limited frequency content, governed by the measurement time range and the diffusive nature of the fields. Using a synthetic example, we show that TEM inversion using the Cole–Cole model suffers from significant instability when the frequency content of the data is insufficient to constrain all four Cole–Cole parameters. The Debye decomposition model provides a way to overcome this challenge by allowing explicit selection of the relaxation time band over which the influence of chargeability in the data behaves distinctly from the influence of resistivity. In this paper, we introduce a workflow for inverting TEM data using a Debye decomposition model. We demonstrate how to set the relaxation time band when inverting TEM data. We then compare the model recovered from the TEM data with a model obtained by inverting spectral IP (SIP) data from the target, which describes how its resistivity varies as a function of frequency, over a wide frequency range. When inverting SIP data, we use the same Debye decomposition model as was used for the TEM data. The recovered IP parameters can be compared within a common parameterization. Using a synthetic example, we demonstrate that the inversion results obtained from the TEM data are consistent with the bandwidth‐limited model obtained from the SIP response. Finally, we present a field application and show that the method recovers IP parameters that are consistent with expected physical property values for seafloor hydrothermal deposits.