244 Geophysics in Geothermal Exploration 8.2 Velocity versus Azimuth (VVAz): a shift detection methodology The misalignments between the azimuthal stacks and the full-stack full-azimuth are very informative, as it can be translated as velocity anomalies (VVAZ). It is also key to correct them, enhancing the stack’s compatibility before comparing their amplitude variations with the azimuth (AVAZ), to get an accurate estimation of such subtle effects. The shift detection is performed on each azimuthal stack (anisotropic) taking a fullstack full-azimuth as reference (isotropic), as described in the previous chapter during the seismic data conditioning for seismic inversion. The resulting dynamic shifts observed in each sector can be understood as velocity anomalies according to this “isotropic” velocity, associated to the full-stack. Different elements must be considered when choosing this parameter: • The window for the shift detection must been put at its lowest as it controls the vertical resolution of the VVAZ anomaly. • The induced interval variations must be computed, to QC their values. The parameter set must be refined using trials and errors to remain in realistic ranges. • In areas where the signal is of low quality, the shift values must tend to zero, which implies no VVAZ effect. The following sequence is proposed to obtain the interval velocities by azimuthal sector: 1. Compute the average velocity from the interval velocities (isotropic). It corresponds to the average of the interval velocities, in TWT domain. 2. Compute correction coefficients by sector: Coef TWT shift Vavg = − 3. As the depth of events is the same, the multiplication of such coefficients with the average velocities (isotropic) leads to corrected average velocities by azimuthal sector. 4. A Dix formula variation allows to estimate the interval velocities by azimuthal sector from the average velocities. At the end of the process, as many interval velocity models as azimuthal stacks are obtained, from which anisotropy can be extracted.
RkJQdWJsaXNoZXIy NjA3NzQ=