Data discrepancy constraints for 3-D airborne transient electromagnetic inversion for induced polarization parameters

Many academics have been interested in the induced polarization (IP) effect in the airborne transient electromagnetic method (ATEM) in recent years. Because most electromagnetic inversion methods only use resistivity parameters and do not contain the IP information of rocks and ores, it is difficult for such methods to retrieve subsurface IP parameters from ATEM data that includes IP effect. The development of unique inversion techniques is becoming increasingly important for obtaining more detailed characterizations of underground-induced polarization. In conventional inversion about IP parameters of ATEM data, only resistivity and polarizability are often fully recovered, while the inversion results of relaxation time and frequency dependence do not reflect any valid information. The data discrepancy constraints based on z-score normalization enable for the compression of data of variable magnitudes to a uniform scale while keeping their trend properties, resulting in a more accurate portrayal of the correlation of trends between different datasets. We assume the IP parameters exhibit a strong correlation among them and use such relationship as the constraint in our 3D inversion algorithm. The inversion method was evaluated on synthetic data, proving that incorporating data discrepancy constraints can improve the usability of inversion results to reveal vital underground IP information, hence simplifying the geological interpretation process. In addition, we used various amounts of noise to assess the resilience of the constrained inversion approach, and the inversion results consistently mirrored the location of the actual target body. As a result, there is significant potential for expanding the data discrepancy constraints to various geophysical inversion techniques.