LAUSR

Induced polarization (IP) effects can distort late-time transient electromagnetic (TEM) signals, sometimes even leading to their sign reversal. These distortions are closely linked to subsurface IP effects. However, the significantly different sensitivities of the IP parameters often result in strong nonuniqueness in TEM data inversion. Here, we present a 3-D TEM-IP inversion method that incorporates local Pearson correlation constraints (LPCCs). By establishing LPCC between the resistivity and the IP parameters, we achieve a joint inversion framework that stabilizes the estimation of these parameters. Numerical experiments on synthetic models demonstrate that, compared with unconstrained inversion, our method yields results that can more accurately recover the true distributions of chargeability, time constant, and frequency-dependent coefficient. We further investigate the impact of the size of the correlation window in the LPCC domain, and achieve practical guidance for optimal constraint selection. Finally, we apply the proposed method to a TEM dataset collected from a bauxite exploration site in Guangxi Province, Southern China, to validate our method for real-world geological settings.