A Robust Interval- Q Inversion Method Based on Q -scanning and Spectral Matching

Abstract Spectral-ratio (SR) methods for estimating the quality factor (Q) from post-stack seismic data often suffer from numerical instability and are highly sensitive to noise. We propose a robust interval-Q inversion method based on Q-scanning and spectral matching. The method consists of three key steps. First, stationary deconvolution is applied to whiten the seismic spectrum and reduce the influence of the source wavelet. Second, instead of using the highly unstable spectral ratio in the frequency domain, the average-Q model is estimated by matching the forward-modelled spectra (within a predefined Q-scanning range) to the whitened frequency-domain data. Third, interval-Q estimation is formulated as an inverse problem. The problem is solved by incorporating time-varying smoothing regularization into the objective function. And a projection-based hard constraint is also imposed to keep Q within the admissible range.The key feature of the proposed framework is the integration of Q-scanning-based spectral matching, which avoids the numerical instability associated with spectral division, with a robust, physically constrained inversion for stabilizing the conversion from average Q to interval Q. Synthetic and field tests demonstrate that the method is less sensitive to noise and produces more continuous and artifact-reduced interval-Q fields than conventional SR-based workflows. These results indicate its potential for attenuation characterization, inverse Q filtering, and reservoir interpretation.