Inversion for Fracture Indicators Combining AVAZ and VVAZ Features in Azimuthal Seismic Data

Seismic waveforms at large offsets are typically stretched during normal moveout (NMO) correction, and as a result, they are often muted in seismic processing. However, in horizontally transverse isotropic (HTI) media, variation of seismic wave reflection with azimuth becomes more pronounced at large offsets. This highlights the importance of retaining large-offset seismic waveforms to achieve better characterization of HTI media and fractured reservoirs. To address the challenges of NMO stretching and preserve seismic waveforms at large offsets, we propose an approach and workflow that utilizes seismic waveforms without NMO correction to estimate fracture weaknesses. We first derive PP-wave reflection coefficient and azimuthal EI (AEI) as functions of fracture weaknesses, and using the derived AEI, we present a novel convolutional model to generate PP-wave seismic data without NMO correction. We then present an inversion method of employing the Bayesian Markov chain Monte Carlo (MCMC) algorithm to estimate unknown parameter vector involving elastic parameters and fracture weaknesses. We use noisy synthetic seismic data to verify the robustness of the proposed inversion method. Applying the proposed inversion method to real data, we obtain reliable inversion results of fracture weaknesses, which provides valuable insights for identifying potential fractured reservoirs.