One-dimensional full-waveform inversion for magnetic induction data in ground-based transient electromagnetic methods

Abstract

The full waveform effects refer to the total effects of turn-on, steady, and turn-off durations for a transmitting-current waveform as well as its repetition number in transient electromagnetic (TEM) methods. In this study, the full waveform effects are investigated using both forward-modelling and inversion methods considering typical background noise. The forward-modelling results of homogeneous half-space models show that the magnetic induction, bz, is less affected by the background noise but more affected by the full waveform effects than the time derivative of magnetic induction ∂bz/∂t. Therefore, this study focuses on investigating the full waveform effects on bz. The inversion results for synthetic and field examples show that the inversion algorithm without considering the full waveform effects leads to over-estimated resistivities in deeper parts of the recovered models compared to the true model. Therefore, it is crucial to consider the full waveform effects when processing TEM data. Furthermore, a standard deviation factor (STDF) is estimated for model parameters of the inversion. The results show that the STDF increases as the layer depth increases for 1D layered models. This indicates that the inverted parameters are well resolved for shallower layers and moderately to poorly resolved for deeper layers.