Fracture prediction using shear-wave splitting in azimuthal common-image gathers

Abstract

Multi-layered fracture systems are common in hydrocarbon reservoirs, and their complex spatial distributions significantly affect reservoir storage capacity and fluid flow. Therefore, accurately predicting key attributes such as fracture strike and intensity is crucial. Converted P-to-S (PS) waves carry information from both P-wave azimuthal anisotropy and shear-wave splitting, offering richer seismic responses for identifying fractures in complex formations. However, predicting fracture parameters from prestack PS-wave data remains challenging because of complex data processing workflows and inversion stability issues. The challenges of PS-wave analysis in multi-layered formations are addressed using a multi-step method. First, an integrated approach combining Alford rotation and layer stripping was utilized to effectively separate fast and slow PS-waves within multi-layered anisotropic systems. Following wavefield separation, the fracture strike was estimated using a weighted circular mean algorithm, and the fracture intensity was quantitatively characterized via sliding-window-based Z-score normalization. This comprehensive workflow enables reliable fracture parameter prediction, even from offshore data with limited azimuthal coverage. Numerical simulations and real-world data applications confirmed a simultaneous enhancement in both the quality of PS-wave imaging and reliability of quantitative fracture characterization in multi-layered formations, providing an effective approach for shear-wave splitting analysis and fracture parameter prediction using PS-waves under challenging geological conditions.