Estimation of the in-situ stress fields from P-wave seismic data- a real data example

Knowledge of the in-situ subsurface stress fields, including their orientations and magnitudes provide the directions along which the most permeable fractures are present and/or are likely to occur during hydraulic fracturing. Thus, their estimation is important for characterizing the naturally fractured, unconventional, and carbon sequestered reservoirs. The stress fields can be measured at the well locations and multicomponent seismic data are effective for estimating them away from the wells. Because acquisition and processing of three-dimensional (3D) multicomponent seismic data is expensive, in most areas single component (P-wave) 3D seismic data are generally available. Consequently, estimating the subsurface in-situ stress fields from traditional 3D P-wave seismic data is of practical importance. Stress fields induce azimuthal anisotropy resulting in distinct differences in the azimuthal reflection amplitudes in the recorded P-wave seismic data. In this work, we combine pre-stack waveform inversion (PWI) with amplitude-variation-with-offset-and azimuth (AVOZ) analysis in estimating the stress fields from the azimuthally dependent P-wave reflection amplitudes. By sorting the recorded data in into different azimuth bins, and then applying PWI, we first estimate accurate velocity fields along different azimuths. Converting offset domain azimuthal pre-stack data into incidence-angles using the PWI velocity fields, we then stack the angle domain gathers to obtain angle stacks for each azimuth. Finally, we apply AVOZ analysis on these angle stacks to estimate the orientations of the subsurface stress fields. Applying the methodology on a real 3D seismic data from the Rock-Springs uplift, Wyoming, USA, we demonstrate that our method can reliably obtain the subsurface stress fields that are consistent with what is observed at a well location. We conclude that our method is an effective tool for estimating the subsurface stress fields from 3D P-wave seismic data.