9. AVO Analysis of Mobil Offshore Data by a Linearized Inversion in the τ – p Domain
Ganyuan Xia, Mrinal Kanti Sen, Paul L. Stoffa · Society of Exploration Geophysicists eBooks · 1998
We developed a linearized inversion scheme in the τ – p domain based on a weighted stack technique, using a modification of an algorithm proposed by Smith and Gidlow (1987). The algorithm assumes that the background velocity model is known and obtains perturbations to the background from the weighted stack. The weights are computed from a linearized approximation of the reflection coefficients. We assume that the background shear-wave velocity follows the mud-rock line, and we use Gardner's equation to relate density to compressional wave velocity. The inversion results in estimates of fractional changes in compressional wave velocity (Δα/α) and shear-wave velocity (Δβ/β). Two other factors, pseudo Poisson's ratio (ΔQ/Q) and a fluid factor (ΔF), are also calculated by a linear combination of the two perturbation parameters. The fluid factor can be used as a diagnostic for direct detection of hydrocarbons. We first applied the weighted stack technique successfully to synthetic data and tested for the sensitivity of the results to the background velocity. Given the correct background velocity, the algorithm recovers the P-wave and S-wave reflectivities and the fluid factor very well. The results are, however, sensitive to the background velocity model. Then we applied the inversion technique to the Mobil AVO data set. We derived four sections from the data—Δα/α, Δβ/β, ΔQ/Q, and ΔF. The fluid factor profile indicates a hydrocarbon (gas) zone at around 1.97 s which correlates well with the well-log data. We performed the inversion using CMP gathers and assuming a 1-D earth model, and we derived the background velocity by interactively perturbing the velocity model until the reflection events were aligned. We believe that a more sophisticated velocity analysis may help in better resolving lateral continuity of parameters.