Sequential Inversion Method in Time–Frequency Domain for Seismic P‐Wave Impedance

P‐wave impedance inversion is a fundamental technique for linking seismic observations to subsurface rock properties. However, due to the band‐limited nature of seismic wavelets and the frequency‐dependent attenuation during propagation, seismic data exhibit frequency‐dependent variations in amplitude, energy distribution and signal‐to‐noise ratio (SNR). Although frequency information has been incorporated in various inversion approaches, these frequency‐dependent characteristics are not always explicitly and systematically utilized, which may limit inversion accuracy and stability. To address this issue, we propose a sequential inversion method in the time–frequency domain for P‐wave impedance estimation. The method employs the generalized S‐transform to obtain a time–frequency representation of seismic data and introduces an adaptive frequency‐dependent regularization coefficient derived from the seismic amplitude spectrum, enabling the inversion to account for SNR variations across frequencies. A multistage inversion strategy is further adopted in which low‐frequency components are sequentially followed by higher frequency components, allowing progressive reconstruction from large‐scale trends to fine‐scale details. Synthetic and field data experiments demonstrate that the proposed method improves both inversion stability and accuracy. For field seismic data, the inversion root‐mean‐square error decreases by about 30% (from 308.64 to 215.03) when using a constant initial model. These results indicate that the proposed method effectively exploits frequency‐dependent information in seismic data and reduces the dependence on the initial model, providing a robust and accurate way for seismic inversion and reservoir characterization.